Keywords: Autonomous systems
Abstract: This talk will provide an overview of research and technology development efforts for controlling uncrewed vehicles operating in the aerial and maritime domains, and across domains. In the first part of the talk, we will describe autonomous control development for a variety of bio-inspired and hybrid uncrewed vehicles including a family of single and double-winged micro air vehicles (MAVs) inspired by fruits and seeds. We will also highlight selected application areas in omni-directional sensing and navigation-aiding. In the second part of the talk, we will discuss control schemes for uncrewed maritime vehicles and demonstrate application to relevant missions. We will describe multi-vehicle control and experiments in coordinated control across domains. The talk will conclude by taking a brief look at human-autonomy teaming in the context of control for optionally-crewed air vehicles.

Keywords: Machine learning, Modeling, Model Validation
Abstract: This paper presents a novel hybrid modeling approach that combines physics-based modeling and machine learning (ML) methods to enhance soil water content dynamics prediction for agricultural decision-making. The proposed approach outperforms current methods for predicting water content based on soil-water physics or purely on data-driven strategies. Initially, a Markov chain-based model is employed to estimate soil water content. The uncertainty associated with the model-based estimation is then quantified by applying ML algorithms, such as support vector machine (SVM), random forest (RF), and feedforward neural network (FNN), to the soil water content data obtained from the Kansas mesoscale network (Mesonet). This quantified uncertainty reveals the complex soil water content dynamics that analytical Markov chain-based models cannot capture. The term physics-informed ML (PIML) refers to integrating soil-water physics principles with the predictive capabilities of data-driven models. Furthermore, multiple Mesonet time series datasets are utilized to assess the influence of physics-based knowledge on ML predictions. The evaluation of the proposed PIML models demonstrates significant improvements in predicting soil water content.

Keywords: Machine learning
Abstract: Transfer learning has seen significant progress in the domains of supervised learning and reinforcement learning, yet there remains a noticeable gap in the area of regression. In supervised learning, various transfer learning methods have been developed to leverage knowledge from one task to improve performance on another, but these approaches are primarily designed for classification tasks. In the realm of reinforcement learning, many techniques focus on policy transfer, and those that do address samples transfer predominantly apply it within homogeneous contexts. This paper presents a novel algorithm that bridges the transfer learning gap between supervised learning and reinforcement learning while specifically addressing the regression problem of dynamical systems' model estimation. Our approach harnesses the feature extraction capabilities of autoencoders and the generative power of Generative Adversarial Networks (GANs) to train a mapping that facilitates the seamless transformation of samples between dynamical systems. This approach represents a significant advancement in the field of transfer learning, as it offers a versatile and effective solution for transferring regression models across heterogeneous domains. Our experimental results demonstrate the algorithm's efficacy and its potential to improve model generalization and adaptability in diverse scenarios with varying data distributions and dynamics.

Keywords: Machine learning, Lyapunov methods, Nonlinear systems identification
Abstract: This paper presents a deep neural network (DNN)- and concurrent learning (CL)-based adaptive control architecture for an Euler-Lagrange dynamic system that guarantees system performance for the first time. The developed controller includes two DNNs with the same output-layer weights to ensure feasibility of the control system. In this work, a Lyapunov-and CL-based update law is developed to update the output-layer DNN weights in real-time; whereas, the inner-layer DNN weights are updated offline using data that is collected in real-time. A Lyapunov-like analysis is performed to prove that the proposed controller yields semi-global exponential convergence to an ultimate bound for the output-layer weight estimation errors and for the trajectory tracking errors.

Keywords: Machine learning, Adaptive control, Stochastic optimal control
Abstract: This paper presents a model-free reinforcement learning (RL) algorithm for the output-difference feedback control (ODFC) of stochastic systems with measurement and process noise. A policy iteration algorithm is presented using a quadratic output-difference reward function to learn the optimal ODFC from noisy output difference measurements. It is proved that the proposed method trains stable dynamic control laws that approach the analytical optimal solution, even in the presence of process and measurement noise. Simulation results are presented to demonstrate the effectiveness of the proposed scheme.

Keywords: Machine learning, Network analysis and control, Emerging control applications
Abstract: In this paper, we study the problem of unsupervised graph representation learning by harnessing the control properties of dynamical networks defined on graphs. Our approach introduces a novel framework for contrastive learning, a widely prevalent technique for unsupervised representation learning. A crucial step in contrastive learning is the creation of `augmented' graphs from the input graphs. Though different from the original graphs, these augmented graphs retain the original graph's structural characteristics. Here, we propose a unique method for generating these augmented graphs by leveraging the control properties of networks. The core concept revolves around perturbing the original graph to create a new one while preserving the controllability properties specific to networks and graphs. Compared to the existing methods, we demonstrate that this innovative approach enhances the effectiveness of contrastive learning frameworks, leading to superior results regarding the accuracy of the classification tasks. The key innovation lies in our ability to decode the network structure using these control properties, opening new avenues for unsupervised graph representation learning.

Keywords: Machine learning, Formal verification/synthesis, Agents-based systems
Abstract: We introduce a decentralized reinforcement learning (RL) algorithm for swarm systems where reward machines (a type of Mealy machines) are used to encode the non-Markovian reward functions. We use the generalized moments (GMs) to characterize swarm features. Each agent estimates the GM of the swarm state and uses it to estimate the state of the reward machine. The agent then uses the estimated state of the reward machine to update its q-values. We use the gossip algorithm for communication between the agents. Agents exploit this communication to update their estimated GM of the swarm state, which leads to an update of their estimated state of the reward machine. We derive an upper bound for the error between the estimate GM of the swarm state and the ground truth GM of the swarm state, and using that upper bound; we prove the convergence of our proposed algorithm, swarm q-learning with reward machines (Swarm-QRM). To demonstrate the efficiency of our approach, we present two case studies wherein, for Case Study 1, a swarm of agents will perform a pickup and delivery task, and in Case Study 2, the swarm of agents will conduct a search and rescue task. We also show that Swarm-QRM outperforms the baselines, q-learning in augmented state space (QAS) and Double Deep Q-Network (DDQN).

Keywords: Adaptive control, Computer-aided control design, Manufacturing systems
Abstract: Over the past three decades, the primary methods for managing complex chemical processes have predominantly relied on PI/PID controllers and model predictive controllers (MPCs). Despite their advancements, both approaches have limitations. PI/PID controllers require specific tuning for each scenario, while MPCs are computationally intensive. To address these challenges, we introduce the LSTM-controller (LSTMc), which is a model-independent, data-driven framework that leverages the robust time-series prediction capabilities of LSTM networks. The LSTMc predicts future control inputs by analyzing the evolution of the system's state and error dynamics, taking into account both the current and previous time steps. Our experiments, particularly in dextrose batch crystallization, have demonstrated the effectiveness of this approach. Impressively, the LSTMc achieves a set-point deviation of less than 2% in 10+ case studies we tested. These results position the LSTMc as a promising alternative for process control, as it adeptly adapts to changing process conditions and set-points, and offers efficient computation for optimal control inputs.

Keywords: Chemical process control, Machine learning, Predictive control for nonlinear systems
Abstract: We investigate the application of state-of-the-art recurrent machine learning methods to tackle the computational challenges associated with the real-time solvability of a packed-bed ammonia synthesis reactor model for optimal control design. A high-fidelity multi-physics model is developed for the ammonia reactor by integrating reaction kinetics and complex transport phenomena. The model involves a system of interconnected algebraic, ordinary, and partial differential equations, providing a deeper understanding and improved prediction accuracy of pressure, temperature, and species concentration dynamics. However, due to its computational complexity, the resulting model is unsuitable for optimal control applications such as model predictive control (MPC). To overcome this limitation, we create a reduced-order model based on long short-term memory (LSTM) using time series data generated from offline simulations of process outputs. This surrogate model effectively captures the system’s nonlinear dynamics, enabling the successful implementation of MPC and achieving optimal closed-loop performance.

Keywords: Fuzzy systems, Optimization, Machine learning
Abstract: Optimization-based decision-making in practical settings often produces solutions that are counter-intuitive to operators and human decision makers. In this paper, we propose an approach for explaining optimization results. A fuzzy inference model based on a set of observations generated by solving an optimization problem is developed, and the logical rules are used to explain the optimization results. The proposed approach is demonstrated using a realistic example of the unit commitment problem.

Keywords: Game theory, Machine learning
Abstract: Hamilton-Jacobi-Isaacs (HJI) PDEs are the governing equations for the two-player general-sum games. Unlike Reinforcement Learning (RL) methods, which are data-intensive methods for learning value function, learning HJ PDEs provide a guaranteed convergence to the Nash Equilibrium value of the game when it exists. However, a caveat is that solving HJ PDEs becomes intractable when the state dimension increases. To circumvent the curse of dimensionality (CoD), physics-informed machine learning methods with supervision can be used and have been shown to be effective in generating equilibrial policies in two-player general-sum games. In this work, we extend the existing work on agent-level two-player games to a two-player swarm-level game, where two sub-swarms play a general-sum game. We consider the Kolmogorov forward equation as the dynamic model for the evolution of the densities of the swarms. Results show that policies generated from the physics-informed neural network (PINN) result in a higher payoff than a Nash Deep Q-Network (Nash DQN) agent and have comparable performance with numerical solvers.

Keywords: Grey-box modeling, Machine learning, Neural networks
Abstract: Due to safety concerns, direct implementation of black-box tools for fully data-driven deep-neural-network (DNN) digital twins faces challenges. To overcome this, hybrid models that combine physics-based first principles (FP) with machine learning (ML) have gained popularity, seen as a `best of both worlds' solution. However, current simple DNN models struggle to predict long-term process data evolution. Recently, time-series transformers (TSTs) using multi-headed attention mechanisms have shown excellent predictive performance for a wide array of time-series prediction tasks. To this end, a novel TST-based hybrid modeling framework for batch crystallization was developed, offering enhanced accuracy and interpretability compared to traditional black-box models. Specifically, the developed TST-based hybrid model allows online estimation of instantaneous growth and nucleation rate (i.e., mathbf{G}&mathbf{B}) for any arbitrarily chosen operating condition in batch crystallization of dextrose. The estimated mathbf{G}&mathbf{B} values can then be directly fed to an FP crystallization model to generate the evolution of crystal size distribution (CSD) and other system states. Moreover, the developed TST-based hybrid model achieved a normalized mean square error in the range of [10, 50]times10^{-4} and an R^2 value exceeding 0.99 for predicting system states. Moreover, the proposed TST-based hybrid model utilizes powerful attention mechanisms to (a) under short and long-term contextual information, and (b) seamlessly learn the functional dependence of system states, thereby setting the foundation stones for next-generation attention-powered hybrid models.

Keywords: Grey-box modeling, Simulation, Machine learning
Abstract: Real-world data of the operation of control systems is often scarce because experiments are expensive and time-consuming. We address the problem of synthetic data generation, namely, the problem of generating data about the operation of a physical system through computational means. In this paper, we report generative adversarial network (GAN) models that can incorporate training data and governing equations underlying the operation of a control system. Instead of over-generalization, we restrict the discussion to the particular example, namely, the Zermelo navigation problem of a vehicle navigating a drift field (e.g., wind) in minimum time. Owing to the nature of the example chosen, we find algebraic governing equations. We propose GAN models that learn to generate trajectories that not only resemble the training data but also satisfy these governing equations. We compare the proposed models to a standard GAN model that uses training data, only, i.e., neglects the governing equations, to demonstrate that the proposed models outperform the standard model on several metrics. To the best of our knowledge, this is the first work on generative models for control systems.

Keywords: Learning, Markov processes, Machine learning
Abstract: Practical applications of reinforcement learning (RL) often demand that the agents explore safety by satisfying designed constraints. The constraints for safety cannot be satisfied with probability 1 when an unbounded stochastic uncertainty is present. One way is to relax the hard constraints into chance constraints and to consider safe RL with chance constraints. This paper addresses safe RL with chance constraints by Model Predictive Control (MPC) with Probabilistic Control Barrier Function (PCBF). MPC with PCBF is used as a function approximator to deliver the policy that satisfies the demanded chance constraints. RL is used to optimize the parameters in MPC with PCBF to improve the closed performance. We prove that using PCBF as a constraint in MPC ensures the safety imposed by the chance constraints. A scenario-based algorithm is designed for the proposed safe RL implementation. A quadrotor system control problem in turbulent conditions has been used as a numerical example to validate the proposed safe RL method.

Keywords: Optimization, Optimization algorithms, Machine learning
Abstract: Delays and asynchrony are inevitable in large-scale machine-learning problems where communication plays a key role. As such, several works have extensively analyzed stochastic optimization with delayed gradients. However, as far as we are aware, no analogous theory is available for min-max optimization, a topic that has gained recent popularity due to applications in adversarial robustness, game theory, and reinforcement learning. Motivated by this gap, we examine the performance of standard min-max optimization algorithms with delayed gradient updates. First, we show (empirically) that even small delays can cause prominent algorithms like Extra-gradient (texttt{EG}) to diverge on simple instances for which texttt{EG} guarantees convergence in the absence of delays. Our empirical study thus suggests the need for a careful analysis of delayed versions of min-max optimization algorithms. Accordingly, under suitable technical assumptions, we prove that Gradient Descent-Ascent (texttt{GDA}) and texttt{EG} with delayed updates continue to guarantee convergence to saddle points for convex-concave and strongly convex-strongly concave settings. Our complexity bounds reveal, in a transparent manner, the slow-down in convergence caused by delays.

Keywords: Neural networks, Identification for control, Machine learning
Abstract: The development of an efficient dynamic model for solid oxide fuel cell (SOFC) systems is essential for control strategies, process optimization, and fault diagnostics - key steps toward maximizing power generation and extending their lifetime. Long-Short-Term-Memory (LSTM), a potent type of recurrent neural network, has proven to be adept at modeling intricate physical dynamic systems using extensive input-output data. They are particularly well-suited for capturing long-range dependencies and temporal patterns. Nevertheless, due to memory and computing constraints in practice, training an LSTM network to effectively learn long-term dependency is challenging, which stems from the sequential processing nature of LSTMs. In this study, a self-supervised convolutional autoencoder (AE) is developed to learn a concise yet informative temporal representation of historical input data. These compressed hidden sequences are then input into an LSTM model, which is trained using a truncated backpropagation through time (TBPTT) algorithm. Once trained, an identical stateful LSTM model is reconstructed using the learned parameters, enabling the LSTM to predict the voltage using only the most recent past input data, rather than a sequence of data over which the model was trained. The efficacy of the proposed framework is validated utilizing diverse experimental data collected from a lab-scale SOFC. The results indicate while the model properly identifies the underlying dynamics of the SOFC, it significantly reduces the training and runtime prediction computation costs of a traditional LSTM model by 38% and 42% respectively. This enhancement holds promise for improving real-time control and diagnostics of SOFC systems, ultimately contributing to their improved performance and reliability.

Keywords: Identification for control, Nonlinear systems identification, Machine learning
Abstract: Electrical Throttle Bodies (ETBs) are massively used in the automotive industry and their modeling and control are challenging because of their high nonlinearity and stochasticity. In this paper, we present a data-driven method grounded on the Koopman operator for the identification of a real ETB valve. The model obtained is control-oriented and represented in a quasi-linear-parameter varying framework. Different experiments are performed to evaluate the performance of the proposed method, including the comparison with two classical nonlinear system identification methods.

Keywords: Optimization, Optimization algorithms, Machine learning
Abstract: In this paper, we propose novel Ising machines hyperparameter tuning techniques for practical use when handling multiple combinatorial problem instances in a short period of time. Firstly, we confirm the performance of a well-known hyperparameter tuning technique, Tree-structured Parzen Estimator (TPE), when targeting a single input instance within a long period of time. Secondly, we propose an efficient method consisting of a generic hyperparameter tuning and a short-time specialized hyperparameter tuning for multiple instances. Since the generic hyperparameter tuning to cover a wide range of instances in a certain category is done in advance, taking time for it is acceptable. The specialized hyperparameter tuning targeting a tuning time of several minutes, reflecting practical use, runs when the performance of a generic hyperparameter set is below expectations. Thirdly, we compare the proposed method to the TPE based tuning to show its effectiveness. For experiments, well-known Travel Salesman Problem (TSP) and Quadratic Assignment Problem (QAP) instances are used as input. The Ising machine used is Fujitsu’s third generation Digital Annealer (DA). Results show that the generic hyperparameter tuning outperforms the TPE based tuning by 2 times in GAP, which indicates the quality of the solution, and the specialized hyperparameter sets have 1.4 times better performance than the generic ones in GAP.

Keywords: Predictive control for nonlinear systems, Kalman filtering, Machine learning
Abstract: In recent years, reinforcement learning (RL) has made substantial progress by providing feasible solutions to many planning and control problems. However, the majority of RL algorithms, particularly model-free RL, suffer from low learning efficiency. To address this issue, this paper proposes the utilization of the Kalman Bayesian neural network (KBNN) to learn a tractable dynamics model of a system from data that captures uncertainties of the system state. Additionally, we employ the unscented Kalman filter to propagate these uncertainties over the control horizon, and we exploit the propagated uncertainties explicitly in the cost function. This approach presents a novel solution to improve data-efficiency in RL. This is validated on classic control problems in a comparative analysis against state-of-the-art methods.

Keywords: Multivehicle systems, Modeling, Machine learning
Abstract: Motion prediction for intelligent vehicles typically focuses on estimating the most probable future evolutions of a traffic scenario. Estimating the gap acceptance, i.e., whether a vehicle merges or crosses before another vehicle with the right of way, is often handled implicitly in the prediction. However, an infrastructure-based maneuver planning can assign artificial priorities between cooperative vehicles, so it needs to evaluate many more potential scenarios. Additionally, the prediction horizon has to be long enough to assess the impact of a maneuver. We, therefore, present a novel long-term prediction approach handling the gap acceptance estimation and the velocity prediction in two separate stages. Thereby, the behavior of regular vehicles as well as priority assignments of cooperative vehicles can be considered. We train both stages on real-world traffic observations to achieve realistic prediction results. Our method has a competitive accuracy and is fast enough to predict a multitude of scenarios in a short time, making it suitable to be used in a maneuver planning framework.

Keywords: Biomolecular systems, Systems biology, Cellular dynamics
Abstract: Chaperones are a global resource within cellular biomolecular systems, ensuring that proteins are properly folded and preventing that aggregation leads to cell death. Introducing genetic circuits to a cell may place a load on these folding resources, resulting in unintended coupling between otherwise independent circuits' behavior. Previous analyses have considered loading effects on other cellular resources --- such as gene expression resources --- but have not included chaperone-enabled folding. In this paper, we model two chaperone modalities, encapsulating two important classes of chaperones as well as their potential interactions. We identify distinct responses that arise from the different architectures which can be either competitive or activating. This work indicates that native cellular chaperones may have built-in control architectures to mitigate loading by an increased demand from chaperone-reliant proteins.

Keywords: Sensor networks, Agents-based systems, Networked control systems
Abstract: In this paper, we propose an opinion-based task allocation strategy for mobile sensor networks to conduct simultaneous multi-source tracking while doing measurement collection and information estimation. Mobile sensors form opinions of the field based on the information obtained by sensors, and opinion dynamics describe how sensors communicate with each other and generate decisions on task assignments for multi-source tracking. Triggering conditions of running opinion dynamics for task allocation are also provided, determining when it is necessary to divide sensors into groups. Simulation results with a 12-sensor network tracking two different sources validate the proposed opinion-based task allocation strategy.

Keywords: Control applications, Process Control, Chemical process control
Abstract: Greenhouses play a pivotal role in achieving food security and paving the way for sustainable urban agriculture. Yet, to optimize crop growth, they demand substantial energy, mainly for climate control and artificial lighting. Given this energy-intensive nature, greenhouses are prime candidates for demand response programs, enhancing the power grid’s flexibility and resilience in urban areas. In this study, we propose a multi-agent deep reinforcement learning (DRL) framework to manage energy in interconnected renewable energy integrated greenhouses. The framework addresses the challenges of fluctuating on-site renewable energy outputs, and interaction with a dynamic electricity tariff grid. Our multi-agent DRL strategy employs an actor-critic algorithm infused with a shared attention mechanism to ensure scalability and performance. We demonstrate the efficacy of our approach using a New York City (NYC) case study with a network of five greenhouses in each borough. We also compare the performance of our multi-agent DRL method to that of the rule-based and the deep deterministic policy gradient algorithms.

Keywords: Control of networks, Network analysis and control, Large-scale systems
Abstract: False information is prevalent on social media platforms, necessitating effective countering methods to combat its spread. We propose an algorithm that reduces the false information spread while preserving the spread of correct information. We model the social media network as a random network of users in which each news item propagates in the network in consecutive cascades. Existing studies suggest that the cascade behaviors of misinformation and correct information are affected differently by user polarization and reflexivity. We show that this difference can be used to alter network dynamics in a way that selectively hinders the spread of misinformation content. To implement these alterations, we introduce an optimization-based probabilistic dropout method that randomly removes connections between users to achieve minimal propagation of misinformation. We test the algorithm's effectiveness using simulated social networks. In our tests, we use both synthetic network structures based on stochastic block models, and natural network structures that are generated using random sampling of data collected from Twitter. The results show that on average the algorithm decreases the cascade size of misinformation content by up to 70% in synthetic network tests and up to 45% in natural network tests while maintaining a branching ratio of at least 1.5 for correct information.

Keywords: Distributed control, Simulation, Multivehicle systems
Abstract: Interactive behavior modeling of multiple agents is an essential challenge in simulation, especially in scenarios when agents need to avoid collisions and cooperate at the same time. Humans can interact with others without explicit communication and navigate in scenarios when cooperation is required. In this work, we aim to model human interactions in this realistic setting, where each agent acts based on its observation and does not communicate with others. We propose a framework based on distributed potential games, where each agent imagines a cooperative game with other agents and solves the game using its estimation of their behavior. We utilize iLQR to solve the games and closed-loop simulate the interactions. We demonstrate the benefits of utilizing distributed imagined games in our framework through various simulation experiments. We show the high success rate, the increased navigation efficiency, and the ability to generate rich and realistic interactions with interpretable parameters. Illustrative examples are available at https://sites.google.com/berkeley.edu/distributed-interacti on

Keywords: Information theory and control, Information technology systems, Network analysis and control
Abstract: Misinformation (e.g., rumors and false information) on social networks poses substantial threats in various domains, underscoring the urgent need for interventions to address misinformation propagation. We present an edge-based optimal control algorithm to minimize the spread of misinformation over social networks. The proposed algorithm preserves the level of information each network node receives while preventing rumors from going viral. We introduce a localized version of the algorithm to enable scaling to large networks. Localization is based on removing any connecting edges between two nodes in the underlying graph that exceed the predefined localization distance in the latent space, whereby the probability of interaction between the two nodes decays exponentially with the distance between them in the latent space. The proposed localized control algorithm compensates for the truncation errors and retains the baseline algorithm's guarantees while reducing computational complexity. Empirical studies applying the baseline algorithm on random spatial networks and its localized counterpart on large-scale social networks extracted from 100 million Twitter messages reduced the ratio of infected users by 8% in synthetic networks and up to 6% on real-world social networks while preserving the level of information each node receives and preventing viral rumors.

Keywords: Iterative learning control, Learning, Variational methods
Abstract: Control of a dynamical system without the knowledge of dynamics is an important and challenging task. Modern machine learning approaches, such as deep neural networks (DNNs), allow for the estimation of a dynamics model from control inputs and corresponding state observation outputs. Such data-driven models are often utilized for the derivation of model-based controllers. However, in general, there are no guarantees that a model represented by DNNs will be controllable according to the formal control-theoretical meaning of controllability, which is crucial for the design of effective controllers. This often precludes the use of DNN-estimated models in applications, where formal controllability guarantees are required. In this proof-of-the-concept work, we propose a control-theoretical method that explicitly enhances models estimated from data with controllability. That is achieved by augmenting the model estimation objective with a controlla- bility constraint, which penalizes models with a low degree of controllability. As a result, the models estimated with the proposed controllability constraint allow for the derivation of more efficient controllers, they are interpretable by the control- theoretical quantities and have a lower long-term prediction error. The proposed method provides new insights on the connection between the DNN-based estimation of unknown dynamics and the control-theoretical guarantees of the solution properties. We demonstrate the superiority of the proposed method in two standard classical control systems with state observation given by low resolution high-dimensional images.

Keywords: Networked control systems, Optimization, Optimization algorithms
Abstract: In this paper, we develop a distributed mixing-accelerated primal-dual proximal algorithm, referred to as MAP-Pro, which enables nodes in multi-agent networks to cooperatively minimize the sum of their nonconvex, smooth local cost functions in a decentralized fashion. The proposed algorithm is constructed upon minimizing a computationally inexpensive augmented-Lagrangian-like function and incorporating a time-varying mixing polynomial to expedite information fusion across the network. The convergence results derived for MAP-Pro include a sublinear rate of convergence to a stationary solution and, under the Polyak-{L}ojasiewics (P-{L}) condition, a linear rate of convergence to the global optimal solution. Additionally, we may embed the well-noted Chebyshev acceleration scheme in MAP-Pro, which generates a specific sequence of mixing polynomials with given degrees and enhances the convergence performance based on MAP-Pro. Finally, we illustrate the competitive convergence speed and communication efficiency of MAP-Pro via a numerical example.

Keywords: Networked control systems, Agents-based systems, Multivehicle systems
Abstract: This paper proposes a consensus controller for multi-agent systems that can guarantee the agents' safety. The controller, built with the idea of output prediction and the Newton-Raphson method, achieves consensus for a class of heterogeneous nonlinear systems. The Integral Control Barrier Function is applied in conjunction with the controller, such that the agents' states are confined within pre-defined safety sets. Due to the dynamically-defined control input, the resulting optimization problem from the barrier function is always a Quadratic Program, despite the nonlinearities that the system dynamics may have. We verify the proposed controller using a platoon of autonomous vehicles modeled by kinematic bicycles. A convergence analysis of the leader-follower consensus under the path graph topology is conducted. Simulation results show that the vehicles achieve consensus while keeping safe inter-agent distances, suggesting a potential in future applications.

Keywords: Networked control systems
Abstract: Clock rigidity theory argues that a specific graph condition can guarantee the solvability of the clock parameters in a time-of-arrival (TOA) based sensor network, where the graph represents the network topology. An auxiliary graph has been proposed to establish the connection between clock rigidity and bearing rigidity, but what is lacking in the literature are methods to leverage the auxiliary graph in the analysis of clock rigidity problems. In this paper, we propose a generalized definition of the auxiliary graph and demonstrate the usage of the auxiliary graph for examining clock configuration genericity in applications. We also study a bearing-based method for examining clock rigidity of a multigraph network, where multiple edges between the same node pairs represent multiple different TOA measurements, extending the families of graphs that can be analyzed using existing clock rigidity theory.

Keywords: Optimization, Optimization algorithms, Networked control systems
Abstract: This paper studies the distributed least-squares optimization problem with differential privacy requirement of local cost functions, for which two differentially private distributed solvers are proposed. The first is established on the distributed gradient tracking algorithm, by appropriately perturbing the initial values and parameters that contain the privacy-sensitive data with Gaussian and truncated Laplacian noises, respectively. Rigorous proofs are established to show the achievable tradeoff between the (ϵ, δ)-differential privacy and the computation accuracy. The second solver is established on the combination of the distributed shuffling mechanism and the average consensus algorithm, which enables each agent to obtain a noisy version of parameters characterizing the global gradient. As a result, the least squares optimization problem can be eventually solved by each agent locally in such a way that any given (ϵ, δ)-differential privacy requirement can be preserved while the solution may be computed with the accuracy independent of the network size, which makes the latter more suitable for large scale distributed least-squares problems. Numerical simulations are presented to show the effectiveness of both solvers.

Keywords: Optimization algorithms, Cooperative control, Autonomous robots
Abstract: We study the problem of multi-agent coordination in unpredictable and partially-observable environments with untrustworthy external commands. The commands are actions suggested to the robots, and are untrustworthy in that their performance guarantees, if any, are unknown. Such commands may be generated by human operators or machine learning algorithms and, although untrustworthy, can often increase the robots’ performance in complex multi-robot tasks. We are motivated by complex multi-robot tasks such as target tracking, environmental mapping, and area monitoring. Such tasks are often modeled as submodular maximization problems due to the information overlap among the robots. We provide an algorithm, Meta Bandit Sequential Greedy (MetaBSG), which enjoys performance guarantees even when the external commands are arbitrarily bad. MetaBSG leverages a meta-algorithm to learn whether the robots should follow the commands or a recently developed submodular coordination algorithm, Bandit Sequential Greedy (BSG) [1], which has performance guarantees even in unpredictable and partially-observable environments. Particularly, MetaBSG asymptotically can achieve the better performance out of the commands and the BSG algorithm, quantifying its suboptimality against the optimal time-varying multi-robot actions in hindsight. Thus, MetaBSG can be interpreted as robustifying the untrustworthy commands. We validate our algorithm in simulated scenarios of multi-target tracking.

Keywords: Sensor networks, Stochastic optimal control, Markov processes
Abstract: The trade-off between communication resources and estimation accuracy is widely considered in sensor networks. In this paper we consider the problem of estimating the trajectory of an event-triggered hidden Markov model, {where the controller decides at each time step whether or not the sensor should sample and transmit a measurement to the estimator}. Adopting a Shannon information-theoretic point of view, we quantify the required communication resources by the entropy of the {transmitted observation sequence}, with a special symbol to denote non-transmission. Furthermore we evaluate the trajectory uncertainty by the conditional entropy of the state sequence given the received observations. Simultaneous minimization of the communication resources and state uncertainty is formulated and solved within a partially observable Markov decision process framework, yielding a threshold policy for triggering transmissions.

Keywords: Chemical process control, Distributed control, Predictive control for nonlinear systems
Abstract: An adaptive community detection-based system decomposition approach is developed for distributed model predictive control (DMPC) of integrated process networks. In this approach, the weighted graph representation of the nonlinear state space model is utilized as the foundation for system decomposition. The most modular decomposition is identified for designing a distributed architecture by applying spectral community detection and energy-based separation techniques. The resulting decomposition evolves as the process network transitions through different operating conditions. Consequently, the distributed architecture and the DMPC design are adjusted to improve closed-loop performance and robustness. To illustrate the efficacy of the proposed approach, a benzene alkylation process benchmark under different operating conditions is used. The results of the simulation studies demonstrate that the decompositions derived through spectral community detection, based on the weighted graph representations, lead to enhanced closed-loop performance and improved computational efficiency compared to the commonly used unweighted hierarchical community-based decompositions.

Keywords: Transportation networks, Control of networks, Intelligent systems
Abstract: Power outages and shortages, among other disruptive events, can markedly diminish the charging efficiency of EV fleets, e.g., e-taxis, and compromise their service quality. This paper aims to address the challenge of coordinating e-taxis amidst such power system disruptions. To understand the extent of the problem, we employ a trace-driven simulation to measure how short-term power failures influence e-taxi service quality. Our observations highlight drops in passenger service in affected areas, pointing to a potential disparity in service quality across various regions of a city. In response, we introduce the Fairness-Aware e-taxi fleet Coordination (FAC) algorithm. FAC monitors the power system disruptions and dynamically switches between two control strategies: one that optimizes city-wide performance during normal operations, and another that emphasizes both service fairness and system performance during power disruptions. We put FAC to the trace-driven evaluation using a comprehensive dataset from an existing e-taxi ecosystem, comprising almost 8,000 taxis and averaging 62,100 taxi trips daily. Our data-driven evaluation shows the effectiveness of our solution in terms of providing fair service quality across regions and enhancing the service quality in the affected regions and a city.

Keywords: Control of networks, Networked control systems, Cooperative control
Abstract: This paper presents a novel approach for resilient distributed consensus in multiagent networks when dealing with adversarial agents and imprecision in states observed by normal agents. Traditional resilient distributed consensus algorithms often presume that agents have exact knowledge of their neighbors' states, which is unrealistic in practical scenarios. We show that such existing methods are inadequate when agents only have access to imprecise states of their neighbors. To overcome this challenge, we adapt a geometric approach and model an agent's state by an `imprecision region' rather than a point in mathbb{R}^d. From a given set of imprecision regions, we first present an efficient way to compute a region that is guaranteed to lie in the convex hull of true, albeit unknown, states of agents. We call this region the emph{invariant hull} of imprecision regions and provide its geometric characterization. Next, we use these invariant hulls to identify a emph{safe point} for each normal agent. The safe point of an agent lies within the convex hull of its emph{normal} neighbors' states and hence is used by the agent to update it's state. This leads to the aggregation of normal agents' states to safe points inside the convex hull of their initial states, or an approximation of consensus. We also illustrate our results through simulations.

Keywords: Multivehicle systems, Cooperative control, Predictive control for nonlinear systems
Abstract: Emerging recovery missions bring the need to recover the parafoil system with the vessel. However, current literature of the parafoil system focuses on the fixed-point landing. This unavoidably hinders the cooperation between the parafoil system and the vessel, which limits the flexibility of payload recovery missions. This paper aims to design the distributed control algorithm for cooperative synchronization between the parafoil system and the vessel. Firstly, the cooperative synchronization process is formulated as a trajectory optimization problem based on the dynamics of the parafoil system and the vessel. To reduce the transmission burden in coordinating the two vehicles, the hierarchical distributed control approach is designed, which incorporates the synchronization point update algorithm to coordinate the two vehicles and the nonlinear model predictive control method to leverage the nonlinear dynamic model for generating the control command. Simulation results demonstrate the effectiveness of the proposed hierarchical distributed model predictive control approach for synchronizing the parafoil and the vessel under disturbances. The findings from this study will benefit the design of cooperative recovery algorithms for future payload recovery missions.

Keywords: Game theory, Agents-based systems, Autonomous systems
Abstract: In this paper, we investigate a multi-agent target guarding problem in which a single defender seeks to capture multiple attackers aiming to reach a high-value target area. In contrast to previous studies, the attackers herein are assumed to be emph{heterogeneous} in the sense that they have not only different speeds but also different weights representing their respective degrees of importance (e.g., the amount of allocated resources). The objective of the attacker team is to jointly minimize the weighted sum of their final levels of proximity to the target area, whereas the defender aims to maximize the same value. Using geometric arguments, we construct candidate equilibrium control policies that require the solution of a (possibly nonconvex) optimization problem. Despite its nonconvex nature, we establish a sufficient condition for this problem to admit a unique local minimum that is also global. Subsequently, we validate the optimality of the candidate control policies using parametric optimization techniques. Lastly, we provide numerical examples to illustrate how cooperative behaviors emerge within the attacker team due to their heterogeneity.

Keywords: Traffic control, Learning, Markov processes
Abstract: Intelligent Traffic Signal Control (TSC) aims to optimize urban traffic management. However, the traditional fixed-cycle intersection signal stands as one of the main factors causing traffic congestion. In addition, with the burgeoning complexity of traffic scenarios and escalating travel demands, TSC becomes a daunting challenge. Recently, the exploration of reinforcement learning in the domain of TSC has emerged as a significant research frontier because it can handle complex scenarios and offers scalability. We follow the framework and propose a multi-agent reinforcement learning (MARL) methodology, in which we incorporate two pivotal enhancements. First of all, to ensure efficacy and stability of training, our method seamlessly combines the Actor-Critic framework with the double Q-network. Secondly, we introduce an advanced adaptive curiosity mechanism to exploit intrinsic rewards founded on the agents' varied action space properties, and hence offers a personalized exploration guide for each agent. Our empirical validation, implemented on the Simulation of Urban Mobility (SUMO) platform, covers both homogeneous grid environment and heterogeneous city setting. Notably, in comparison with traditional cyclic scheduling methods and existing advanced MARL solutions, our algorithm exhibited better traffic flow efficiency and robustness.

Keywords: Automotive systems, Autonomous systems, Simulation
Abstract: In this study, we tackle the XYZ-motion planning problem for autonomous vehicles equipped with active suspension systems. Using perception data, we derive two road surface approximations: the encoded road and the estimated road. Using the estimated road, we first form a nonlinear optimization problem to obtain an XYZ-motion plan. For practical applicability, this optimization problem is split into two steps. First, generating a XY path by using the information of soft and hard obstacles. Second, generating a vertical motion plan with the XY path and estimated road. We have tested our methodologies on two examples, where in the first example we have used a synthetic road to demonstrate online XYZ-motion planning and in the second example we have used real-world road surface data to demonstrate offline XYZ-motion planning. We have found that our algorithms give satisfactory results in both the test cases.

Keywords: Autonomous robots, Formal verification/synthesis, Robust control
Abstract: Signal Temporal Logic (STL) is a formal language over continuous-time signals (such as trajectories of a multi-agent system) that allows for the specification of complex spatial and temporal system requirements (such as staying sufficiently close to each other within certain time intervals). To promote robustness in multi-agent motion planning with such complex requirements, we consider motion planning with the goal of maximizing the temporal robustness of their joint STL specification, i.e. maximizing the permissible time shifts of each agent's trajectory while still satisfying the STL specification. Previous methods presented temporally robust motion planning and control in a discrete-time optimization scheme. In contrast, we parameterize the trajectory by continuous Bézier curves, where the curvature and the time-traversal of the trajectory are parameterized individually. We show an algorithm generating continuous-time temporally robust trajectories and prove soundness of our approach. Moreover, we empirically show that our parametrization realizes this with a considerable speed-up compared to state-of-the-art methods based on constant interval time discretization.

Keywords: Autonomous robots, Estimation, Intelligent systems
Abstract: With the rising prominence of WiFi in common spaces, efforts have been made in the robotics community to take advantage of this fact by incorporating WiFi signal measurements in indoor SLAM (Simultaneous Localization and Mapping) systems. SLAM is essential in a wide range of applications, especially in the control of autonomous robots. This paper describes recent work in the development of WiFi-based localization and addresses the challenges currently faced in achieving WiFi-based geometric mapping. Inspired by the field of research into k-visibility, this paper presents the concept of inverse k-visibility and proposes a novel algorithm that allows robots to build a map of the free space of an unknown environment, essential for planning, navigation, and avoiding obstacles. Experiments performed in simulated and real-world environments demonstrate the effectiveness of the proposed algorithm.

Keywords: Mean field games, Cooperative control
Abstract: This paper proposes mean field models to maintain an accurate line-of-sight and tracking between transceivers mounted in mobile unmanned aerial vehicles (UAVs) platforms in the presence of underlying mechanical vibration effects. We describe a two-way optical link beam tracking control that coordinates mobile UAVs in a network architecture to provide reliable network structure, distributed connectivity, and communicability, enhancing terrestrial public safety communication systems. We derive the optical transceiver trajectory tracking problem in which each agent dynamic and cost function is coupled with other optical beam transceiver agent states via a mean field term. We propose two optimal mean field beam tracking control frameworks through decentralized and centralized strategies in which the optical transceivers compete to reach a Nash equilibrium and cooperate to attain a social optimum, respectively. The solutions of these strategies are derived from forward-backward ordinary differential equations and rely on the linearity Hamilton-Jacobi-Bellman Fokker-Planck equations and stochastic maximum principle. Moreover, we numerically compute the solution pair of the resulting joint equations using Newton and fixed point iteration methods to verify the existence and uniqueness of the equilibrium and social optimum.

Keywords: Autonomous robots
Abstract: Safety filters based on control barrier functions (CBFs) have become a popular method to guarantee safety for uncertified control policies, e.g., as resulting from reinforcement learning. Here, safety is defined as staying in a pre-defined set, the safe set, that adheres to the system's state constraints, e.g., as given by lane boundaries for a self-driving vehicle. In this paper, we examine one commonly overlooked problem that arises in practical implementations of continuous-time CBF-based safety filters. In particular, we look at the issues caused by discrete-time implementations of the continuous-time CBF-based safety filter, especially for cases where the magnitude of the Lie derivative of the CBF with respect to the control input is zero or close to zero. When overlooked, this filter can result in undesirable chattering effects or constraint violations. In this work, we propose three mitigation strategies that allow us to use a continuous-time safety filter in a discrete-time implementation with a local relative degree. Using these strategies in augmented CBF-based safety filters, we achieve safety for all states in the safe set by either using an additional penalty term in the safety filtering objective or modifying the CBF such that those undesired states are not encountered during closed-loop operation. We demonstrate the presented issue and validate our three proposed mitigation strategies in simulation and on a real-world quadrotor.

Keywords: Autonomous systems, Predictive control for nonlinear systems, Cooperative control
Abstract: In this paper, we propose a multi-UAV formation control system in obstacle-filled environments. Our design consists of two key components, which are the formation path planning algorithm and the model predictive control (MPC) for quadrotor agents. The proposed path planning algorithm generates a safe reference path for the formation center and secure waypoints for each agent to minimize pattern deformation. The designed MPC controller helps agents track the reference trajectory and maintain formation using data from neighboring agents. Our system also processes sensor data and models surrounding obstacles for the environment instead of only considering the scenarios with a known environment. Several simulation examples are provided to demonstrate the feasibility and effectiveness of the proposed design.

Keywords: Autonomous systems, Stochastic optimal control, Intelligent systems
Abstract: Understanding the intention of vehicles in the surrounding traffic is crucial for an autonomous vehicle to successfully accomplish its driving tasks in complex traffic scenarios such as highway forced merging. In this paper, we consider a behavioral model that incorporates both social behaviors and personal objectives of the interacting drivers. Leveraging this model, we develop a receding-horizon control-based decision-making strategy, that estimates online the other drivers' intentions using Bayesian filtering and incorporates predictions of nearby vehicles' behaviors under uncertain intentions. The effectiveness of the proposed decision-making strategy is demonstrated and evaluated based on simulation studies in comparison with a game theoretic controller and a real-world traffic dataset.

Keywords: Autonomous systems, Multivehicle systems, Optimal control
Abstract: As autonomous vehicles (AVs) become increasingly prevalent, their interaction with human drivers presents a critical challenge. Current AVs lack social awareness, causing behavior that is often awkward or unsafe. To combat this, social AVs, which are proactive rather than reactive in their behavior, have been explored in recent years. With knowledge of robot-human interaction dynamics, a social AV can influence a human driver to exhibit desired behaviors by strategically altering its own behaviors. In this paper, we present a novel framework for achieving human influence. The foundation of our framework lies in an innovative use of control barrier functions to formulate the desired objectives of influence as constraints in an optimal control problem. The computed controls gradually push the system state toward satisfaction of the objectives, e.g. slowing the human down to some desired speed. We demonstrate the proposed framework’s feasibility in a variety of scenarios related to car-following and lane changes, including multi-robot and multi-human configurations. In two case studies, we validate the framework’s effectiveness when applied to the problems of traffic flow optimization and aggressive behavior mitigation. Given these results, the main contribution of our framework is its versatility in a wide spectrum of influence objectives and mixed-autonomy configurations.

Keywords: Autonomous systems, Statistical learning, Autonomous robots
Abstract: We study the problem of data-driven monitoring using an autonomous search team. We consider a typical monitoring task of classifying a search environment into interesting and uninteresting regions as quickly as possible using a search team comprised of mobile sensors (e.g., drones) and mobile charging stations (e.g., ground vehicles). For widespread deployment in the physical world, the search team must also accommodate noisy data collected by the mobile sensors, overcome energy constraints on the mobile sensors which limit their range, and ensure collision avoidance for the charging stations. We address these challenges using a novel, bi-level approach for the monitoring task, where a high-level planner uses past data to determine the potential regions of interest for the drones to visit, and a low-level path planner coordinates the paths for the entire search team to visit these regions subject to the posed constraints. We design the high-level planner using a multi-armed bandit framework. For the low-level planner, we propose two approaches: an optimal integer program-based motion planner and a real-time implementable graph-based heuristic planner. We characterize several theoretical properties of the proposed approaches, including anytime guarantees, upper bounds on computing time, and task completion time. We show the efficacy of our approach in simulations, including one in Gazebo where we identify harvest-ready trees using an autonomous heterogeneous search team.

Keywords: Autonomous systems, Sensor networks, Optimization
Abstract: We address path-planning for a mobile agent to navigate in an unknown environment with minimum exposure to a spatially and temporally varying threat field. The threat field is estimated using pointwise noisy measurements from a sensor network separate from the mobile agent. For this problem, we present a new metric for optimal sensor placement that quantifies reduction in uncertainty in the path cost, rather than the environment state. This metric, which we call the context-relevant mutual information (CRMI), couples the sensor placement and path-planning problem. We propose also an iterative coupled sensor configuration and path planning (CSCP) algorithm. At each iteration, the algorithm places sensors to maximize CRMI, updates the threat estimate using new measurements, and recalculates the path with minimum expected exposure to the threat. The iterations converge when the path cost variance, which is an indicator of risk, reduces below a desired threshold. Through numerical simulations, we demonstrate that the principal advantage of this algorithm is that near-optimal low-variance paths are achieved using far fewer sensor measurements as compared to a standard decoupled method.

Keywords: Constrained control, Autonomous robots, Stability of nonlinear systems
Abstract: This paper addresses the challenge of safe navigation for rigid-body mobile robots in dynamic environments. We introduce an analytic approach to compute the distance between a polygon and an ellipse, and employ it to construct a control barrier function (CBF) for safe control synthesis. Existing CBF design methods for mobile robot obstacle avoidance usually assume point or circular robots, preventing their applicability to more realistic robot body geometries. Our work enables CBF designs that capture complex robot and obstacle shapes. We demonstrate the effectiveness of our approach in simulations highlighting real-time obstacle avoidance in constrained and dynamic environments for both mobile robots and multi-joint robot arms.

Keywords: Air traffic management, Autonomous systems, Constrained control
Abstract: We study the problem of designing safe trajectories for aircraft approach management. Our tractable method designs aircraft trajectories that 1) use only limited admissible maneuvers near the airport, 2) maintains a user-specified separation between aircraft during the entire duration of the approach, and 3) minimizes deviations from user-specified times of arrival at the airport. We use a first-come, first-serve framework to design the trajectories for multiple aircraft by solving a collection of single-aircraft trajectory planning problems. We ensure the safety of the overall system by imposing reachability-based constraints on each planning problem. We identify the constraints as well as the trajectories for aircraft using dynamic programming in a three-dimensional space. We validate the efficacy and safety of our method using historical data from Japan's Haneda International Airport.

Keywords: Control applications, Autonomous robots, Optimal control
Abstract: This paper introduces an experimental platform designed for the validation and demonstration of a novel class of Control Barrier Functions (CBFs) tailored for Unmanned Ground Vehicles (UGVs) to proactively prevent collisions with kinematic obstacles by integrating the concept of collision cones. While existing CBF formulations excel with static obstacles, extensions to torque/acceleration-controlled unicycle and bicycle models have seen limited success. Conventional CBF applications in such nonholonomic UGV models have demonstrated control conservatism, particularly in scenarios where steering/thrust control was deemed infeasible. Drawing inspiration from collision cones in path planning, we present a pioneering C3BF formulation ensuring theoretical safety guarantees for such models. The core premise revolves around aligning the obstacle's velocity away from the vehicle, establishing a constraint to perpetually avoid vectors directed towards it. This control methodology is rigorously validated through simulations and experimental verification on the Copernicus mobile robot (Unicycle Model) and FOCAS-Car (Bicycle Model).

Keywords: Estimation, Autonomous systems
Abstract: This paper address the problem of path deconfliction and collision avoidance for unmanned aerial vehicles equipped with a monocular optical camera. The measurements observed from the camera do not include range, but they do allow for calculation of time-to-collision. In this paper we exploit the idea that a single time-to-collision estimate may have come from any number of intruders with different range and velocities, to define a family of potential intruders. The family of potential intruders is represented by a set of particles that can then be propagated forward in time to represent the set of all potential collision scenarios. A path planning algorithm is then introduced to minimize the collision risk. The method is illustrated with simulation results.

Keywords: Flight control, Mechanical systems/robotics, Mechatronics
Abstract: Miniature autonomous blimps (MABs) have advantages of safe flight and long flight duration. This paper presents a flight controller design for an underactuated MAB to achieve and maintain the inverted pose. When the MAB starts from a bottom-heavy pose, an energy-shaping controller is engaged to achieve a swing-up motion for the MAB to reach the proximity of the inverted pose. Once the criteria for controller switching are met, a stabilizing controller is employed to maintain the inverted pose. The performance of the proposed controller is verified through the experimental results collected from the GT-MAB.

Keywords: Hybrid systems, Switched systems, Autonomous systems
Abstract: Control Barrier Functions (CBF) have provided a very versatile framework for the synthesis of safe control architectures for a wide class of nonlinear dynamical systems. Typically, CBF-based synthesis approaches apply to systems that exhibit nonlinear – but smooth – relationship in the state of the system and linear relationship in the control input. In contrast, the problem of safe control synthesis using CBF for hybrid dynamical systems, i.e., systems which have a discontinuous relationship in the system state, remains largely unexplored. In this work, we build upon the progress on CBF- based control to formulate a theory for safe control synthesis for hybrid dynamical systems. Under the assumption that local CBFs can be synthesized for each mode of operation of the hybrid system, we show how to construct CBF that can guarantee safe switching between modes. The end result is a switching CBF-based controller which provides global safety guarantees. The effectiveness of our proposed approach is demonstrated on two simulation studies.

Keywords: Mechanical systems/robotics, Autonomous systems, Machine learning
Abstract: We employ sequences of high-order motion primitives for efficient online trajectory planning, enabling competitive racecar control even when the car deviates from an offline demonstration. Dynamic Movement Primitives (DMPs) utilize a target-driven non-linear differential equation combined with a set of perturbing weights to model arbitrary motion. The DMP’s target-driven system ensures that online trajectories can be generated from the current state, returning to the demonstration. In racing, vehicles often operate at their handling limits, making precise control of acceleration dynamics essential for gaining an advantage in turns. We introduce the Acceleration goal (Acc. goal) DMP, extending the DMP’s target system to accommodate accelerating targets. When sequencing DMPs to model long trajectories, our Acc. goal DMP explicitly models acceleration at the junctions where one DMP meets its successor in the sequence. Applicable to DMP weights learned by any method, the proposed DMP generates trajectories with less aggressive acceleration and jerk during transitions between DMPs compared to second-order DMPs. Our proposed DMP sequencing method can recover from trajectory deviations, achieve competitive lap times, and maintain stable control in autonomous vehicle racing within the high-fidelity racing game Gran Turismo Sport. Video available: https://sites.google.com/berkeley.edu/racingdmp/home

Keywords: Game theory, Autonomous systems, Optimal control
Abstract: When interacting with other non-competitive decision-making agents, it is critical for an autonomous agent to have inferable behavior: Their actions must convey their intention and strategy. For example, an autonomous car's strategy must be inferable by the pedestrians interacting with the car. We model the inferability problem using a repeated bimatrix Stackelberg game with observations where a leader and a follower repeatedly interact. During the interactions, the leader uses a fixed, potentially mixed strategy. The follower, on the other hand, does not know the leader's strategy and dynamically reacts based on observations that are the leader's previous actions. In the setting with observations, the leader may suffer from an inferability loss, i.e., the performance compared to the setting where the follower has perfect information of the leader's strategy. We show that the inferability loss is upper-bounded by a function of the number of interactions and the stochasticity level of the leader's strategy, encouraging the use of inferable strategies with lower stochasticity levels. As a converse result, we also provide a game where the required number of interactions is lower bounded by a function of the desired inferability loss.

Keywords: Multivehicle systems, Sensor networks, Autonomous robots
Abstract: We consider the problem of routing a team of energy-constrained Unmanned Aerial Vehicles (UAVs) to deploy sensors for monitoring a surface-level phenomenon in the presence of stochastic wind disturbances. Most of the environmental monitoring work considers the case where the sensors and their carrier are on one platform, and the sensing location can be controlled by controlling the carrier vehicle's position. In contrast, airdropping the sensors onto the ground can introduce stochasticity in the sensors' landing (i.e., measurement) locations. We focus on carefully choosing where to drop the sensors so that despite the measurement locations' stochasticity, we can still effectively monitor the surface-level phenomenon. Specifically, we introduce this problem where a team of UAVs must drop sensors within an allotted time budget. The objective is to maximize the mutual information between the phenomenon at Points of Interest and the measurements obtained at the random sensing locations. This is a variant of the Submodular Team Orienteering Problem with an additional constraint on the number of sensors each UAV can carry. We show that the stochasticity in the measurement location makes objective computationally expensive to compute. We propose a surrogate objective with a closed-form expression and an algorithm for the routing problem. We validate the proposed approach through extensive simulations using synthetic and a real-world dataset of Chlorophyll density from a Pacific Ocean sub-region.

Keywords: Optimization, Control applications, Autonomous robots
Abstract: We propose a motion planner for quadrotor unmanned aerial vehicles (UAVs) in windy environments, where the motion is defined by a sequence of Bezier curves in the flat output space of the UAV. The real-time implementable planner incorporates wind information from high-fidelity computational fluid dynamics simulations performed offline and utilizes convexity properties of Bezier curves to enable real-time implementations. For this purpose, we: (i) identify a model for the UAV-wind interaction; (ii) use the OpenFoam software to compute a model of the wind speeds subject to world geometry and boundary conditions; (iii) describe a method for regressing this wind model into a more compact representation; and finally (iv) demonstrate how this representation is amenable to minimum-snap motion planning of quad-rotor UAVs in realistic environments. We validate our approach using simulations and hardware experiments, and show a significant improvement in the thrust used by the UAV in presence of strong winds.

Keywords: Modeling, Neural networks, Process Control
Abstract: In the context of real-time optimization and model predictive control of industrial systems, machine learning, and neural networks represent cutting-edge tools that hold promise for enhancing dynamic modeling. This work presents a transformer neural network architecture for real-time optimization and model predictive control. This network design includes a modified attention mechanism inspired by positional embedding attention from vision transformers and task-specific modifications to the input-output structure of the transformer’s decoder stack. Experiments were conducted using data from a 450 MW coal-fired power plant to evaluate this approach's effectiveness. The transformer neural network was compared with conventional recurrent models, including GRU and LSTM neural networks. The transformer exhibited a 6% increase in the R-squared value of predictions and an 83% reduction in mean squared error. Computation time was also reduced by 84% compared to conventional recurrent models.

Keywords: Modeling, Hybrid systems, Automata
Abstract: This paper proposes a transition system abstraction framework for neural network dynamical system models to enhance the model interpretability, with applications to complex dynamical systems such as human behavior learning and verification. To begin with, the localized working zone will be segmented into multiple localized partitions under the data-driven Maximum Entropy (ME) partitioning method. Then, the transition matrix will be obtained based on the set-valued reachability analysis of neural networks. Finally, applications to human handwriting dynamics learning and verification are given to validate our proposed abstraction framework, which demonstrates the advantages of enhancing the interpretability of the black-box model, i.e., our proposed framework is able to abstract a data-driven neural network model into a transition system, making the neural network model interpretable through verifying specifications described in Computational Tree Logic (CTL) languages.

Keywords: Constrained control, Identification for control, Nonlinear output feedback
Abstract: This paper addresses the challenge of integrating explicit hard constraints into the control barrier function (CBF) framework for ensuring safety in autonomous systems, including robots. We propose a novel data-driven method to derive CBFs from these hard constraints in practical scenarios. Our approach assumes that the forward invariant safe set is either a subset or equal to the constrained set. The process consists of two main steps. First, we randomly sample states within the constraint boundaries and identify inputs meeting the time derivative criteria of the hard constraint; this iterative process converges using the Jaccard index. Next, we formulate CBFs that enclose the safe set using the sampled boundaries. This enables the creation of a control-invariant safe set, approaching the maximum attainable level of control invariance. This approach, therefore, addresses the complexities posed by complex autonomous systems with constrained control input spaces, culminating in a control-invariant safe set that closely approximates the maximal control invariance set.

Keywords: Aerospace, Linear parameter-varying systems, Modeling
Abstract: This paper presents a Physics-Data-Hybrid (PDH) approach to model the flight dynamics of tilt-rotor Vertical-TakeOff-and-Landing (VTOL) aircraft within the transition phase. The tilting rotors enable more agile flight by varying vectored thrusts; however, they render more complex aircraft dynamics and other challenges for modeling. The proposed PDH modeling approach starts from fundamental, nonlinear physical laws, and expresses them in a Linear Parameter-Varying (LPV) representation, which utilizes scheduling parameters to obtain kernels that encode the intrinsic nonlinearity. After that, flight data snapshots during the transition phase are used to determine the effective state space matrices, bypassing the laborious efforts to identify accurate flight coefficients. The proposed PDH approach is validated and its modeling accuracy is assessed via simulations on a tilt-rotor VTOL aircraft. The simulation first collects data snapshots by feeding a control input sequence to excite the aircraft dynamics and the PDH model is then derived. The state responses of the resulting PDH model are then compared against the true transition trajectory. The simulation results demonstrate the PDH model's excellent modeling accuracy.

Keywords: Automotive control, Robust control, Modeling
Abstract: This paper is concerned with the problem of disturbance propagation in interconnected systems with double-integrator open-loop dynamics (e.g. acceleration-controlled automated vehicles). We consider a symmetric bidirectional linear interconnection to control the vehicles, in which each vehicle is interconnected solely to one immediate predecessor and to one immediate follower with the same control gain in these two directions. We prove that in this setting, regardless of the choice of stabilizing controller, it is not possible to keep the displacements between the vehicles bounded, and they grow unbounded depending on the number of vehicles N. In this case, we call the system string unstable. We show that this impossibility of achieving string stability remains under various boundary conditions for the leading vehicle, i.e., whether this is a virtual or physical vehicle.

Keywords: Grey-box modeling, Identification for control, Control applications
Abstract: A physically motivated heater model for precise gas temperature control for use in fuel cell stack test beds is proposed. Additionally, a parameter identification method based on modulating functions is provided. The applicability of the identification method and the model is validated through simulations and real-world experiments. Lastly, a time-optimal controller based on the identified model is evaluated on a fuel cell stack test bed.

Keywords: Lyapunov methods, Feedback linearization, Stability of nonlinear systems
Abstract: Today, electric scooter is a trendy personal mobility vehicle. The rising demand and opportunities attract ride-share services. A common problem of such services is abandoned e-scooters. An autonomous e-scooter capable of moving to the charging station is a solution. This paper focuses on maintaining balance for these riderless e-scooters. The paper presents a nonlinear model for an e-scooter moving with simultaneously varying speed and steering. A PD and a feedback-linearized PD controller stabilize the model. The stability analysis shows that the controllers are ultimately bounded even with parameter uncertainties and measurement inaccuracy. Simulations on a realistic e-scooter with a general demanding path to follow verify the ultimate boundedness of the controllers. In addition, the feedback-linearized PD controller outperforms the PD controller because it has narrower ultimate bounds. Future work focuses on experiments using a self-balancing mechanism installed on an e-scooter.

Keywords: Identification for control, Mechanical systems/robotics, Mechatronics
Abstract: Multi-rotor actuator dynamics modeling spans from idealized to complex representations, often simplifying actuator nonlinearity. In the context of aerial manipulation, precise control is crucial. To address this, we introduce a method for nonlinear model identification, incorporating an adaptable input definition and a variable-resolution RPM measurement algorithm. Our study addresses challenges posed by commercial ESCs designed for manual control. We identify model parameters using small perturbation models, correlating them with the full nonlinear model.

The paper encompasses an overview of the experimental setup, measurement algorithm, input definition, parameter identification, and validation.

Keywords: Modeling
Abstract: During the COVID-19 pandemic, large-scale nucleic acid testing and quarantine were widely adopted as measures to contain the disease. However, it also aroused some controversy whether massive testing lead to increased interpersonal contacts and thus exacerbate the epidemic spreading. In this paper, we study a scalar SIQRS (Susceptible-Infected-Quarantines-Recovered-Susceptible) model incorporating infection, curing, quarantine and vaccination processes. To characterize the effects of massive testings, we assume that the interpersonal contact frequency between individuals increases linearly with the quarantine rate, which is in turn determined by the frequency of massive testings. Via equilibrium and stability analysis, we study the effects of the quarantine rate and find that there exists a threshold, above which massive testings are helpful in containing the disease and below which massive testings facilitate the epidemic spreading. Numerical simulations are conducted to demonstrate the correctness of the theoretical analysis and the effectiveness of the proposed quarantine strategy.

Keywords: Nonlinear systems identification, Estimation, Reduced order modeling
Abstract: Incorporating prior knowledge into a data-driven modeling problem can drastically improve performance, reliability, and generalization outside of the training sample. The stronger the structural properties, the more effective these improvements become. Manifolds are a powerful nonlinear generalization of Euclidean space for modeling finite dimensions. When additionally assuming that the manifold carries (Lie) group structure, this imposes a drastically stricter global constraint. The range of their applications is very wide and includes the important case of robotic tasks. We apply this idea to Canonical Correlation Analysis (CCA). In traditional CCA one constructs a hierarchical sequence of maximal correlations of up to two paired data sets in Euclidean spaces. We here generalize the CCA concept to respect the structure of Lie groups and demonstrate its efficacy through the substantial improvements it achieves in making structure-consistent predictions about changes in the state of a robotic hand.

Keywords: Reduced order modeling, Hierarchical control, Autonomous robots
Abstract: The Angular-Momentum Linear Inverted Pendulum (ALIP) model is a promising motion planner for bipedal robots. However, it relies on two assumptions: (1) the robot has point-contact feet or passive ankles, and (2) the angular momentum around the center of mass, known as centroidal angular momentum, is negligible. This paper addresses the question of whether the ALIP paradigm can be applied to more general bipedal systems with complex foot geometry (e.g., flat feet) and nontrivial torso/limb inertia and mass distribution (e.g., non-centralized arms). In such systems, the dynamics introduce non-negligible centroidal momentum and contact wrenches at the feet, rendering the assumptions of the ALIP model invalid. This paper presents the ALIP planner for general bipedal robots with non-point-contact feet through the use of a task-space whole-body controller that regulates centroidal momentum, thereby ensuring that the robot's behavior aligns with the desired template dynamics. To demonstrate the effectiveness of our proposed approach, we conduct simulations using the Sarcos Guardian XO robot, which is a hybrid humanoid/exoskeleton with large, offset feet. The results demonstrate the practicality and effectiveness of our approach in achieving stable and versatile bipedal locomotion.

Keywords: Estimation, Filtering, Sensor fusion
Abstract: We consider the problem of evaluating dynamic consistency in discrete time probabilistic filters that approximate stochastic system state densities with Gaussian mixtures. Dynamic consistency means that the estimated probability distributions correctly describe the actual uncertainties. As such, the problem of consistency testing naturally arises in applications with regards to estimator tuning and validation. However, due to the general complexity of the density functions involved, straightforward approaches for consistency testing of mixture-based estimators have remained challenging to define and implement. This paper derives a new exact result for Gaussian mixture consistency testing within the framework of normalized deviation squared (NDS) statistics. It is shown that NDS test statistics for generic multivariate Gaussian mixture models exactly follow mixtures of generalized chi- square distributions, for which efficient computational tools are available. The accuracy and utility of the resulting consistency tests are numerically demonstrated on static and dynamic mixture estimation examples.

Keywords: Estimation, Network analysis and control, Simulation
Abstract: Tor is a widely used anonymous communication network that routes and anonymizes users’ internet traffic through thousands of relays. To create a path, Tor authorities estimate relay capacities based on the bandwidth of observation probes assigned to each relay. These estimates are used to generate a probability distribution over relays for incoming users to choose from. However, the currently implemented estimation algorithm generates inaccurate estimates, resulting in underutilization of the network and unfair distribution of capacities between users. To address this, we propose DiProber, a new algorithm that uses two probes per relay and maximum likelihood to more accurately estimate relay capacities. Our new technique works particularly well in underutilized networks where users have low demand on the Tor network.

Keywords: Estimation, Kalman filtering, Filtering
Abstract: This paper presents a novel distribution-agnostic Wasserstein distance-based estimation framework. The goal is to determine an optimal map combining prior estimate with measurement likelihood such that posterior estimation error optimally reaches the Dirac delta distribution with minimal effort. The Wasserstein metric is used to quantify the effort of transporting from one distribution to another. We hypothesize that minimizing the Wasserstein distance between the posterior error and the Dirac delta distribution results in optimal information fusion and posterior state uncertainty. Framework validation is demonstrated by the successful recovery of the classical Kalman filter for linear systems with Gaussian uncertainties. Notably, the proposed Wasserstein filter does not rely on particle representation of uncertainty. Furthermore, the classical result for the Gaussian Sum Filter (GSF) is retrieved from the Wasserstein framework. This approach analytically exhibits the suboptimality of GSF and enables the use of nonlinear optimization techniques to enhance the accuracy of the Gaussian sum estimator.

Keywords: Energy systems, Estimation
Abstract: As surplus of second-life (SL) batteries becomes available, one critical obstacle between their widespread adoption is the accurate estimation and monitoring of their state-of-health (SOH). Various retired battery packs with a lack of knowledge of their historical usage are used to set up new SL battery systems. The paper presents a new online adaptive health estimation method designed to address this practical challenges associated with SL battery systems. This method utilizes the real-time data obtained from the SL batteries used for grid energy storage systems. The proposed approach is validated on a laboratory-aged experimental data set of retired EV batteries. Through dynamic adjustment of estimator gains, the approach can effectively accommodate the unique characteristics of individual cells, enhancing its adaptability and robustness. The bounded-input-bounded-output (BIBO) stability of this adaptive estimator is theoretically guaranteed.

Keywords: Intelligent systems, Adaptive control, Optimization
Abstract: Model-Free control is a data-driven control methodology that utilizes a surrogate model, namely the ultra-local mode, to provide a short-time estimate of the dynamics and uncertainties related to a nonlinear system. Traditional Model-Free control requires two parameters, one is estimated online while the other is considered a tuning parameter for the performance of the system at hand. The tuning parameter is typically chosen heuristically and depends upon expert knowledge of the system. This paper provides a methodology for on-the-fly simultaneous selection of all parameters of the ultra-local model. The procedure relies on a robust Kalman Filter and an optimization routine that enforces the validity of the ultra-local model. An illustrative example is provided where the methodology developed is shown to improve the tracking performance of a system and, more importantly, that the estimated parameters converge to the expected values of the system.

Keywords: Multivehicle systems, Distributed control, Estimation
Abstract: In this paper, we investigate the problem of active joint localization and target tracking of mobile robots with onboard sensors. Our primary objective is concurrent target tracking while precisely localizing the robots through coordinated motion. A key constraint is the distributed setting, where each robot’s observations are limited to its immediate vicinity, and communication is restricted to neighboring robots. To address this, we propose a novel reinforcement learning-based approach for active motion planning, grounded in a distributed estimation framework called Joint Localization and Target Tracking (JLATT). The policy is trained to optimize robot coordination and trajectories for enhanced self-localization, target tracking, and collision avoidance. Empirical analysis demonstrates our algorithm’s effectiveness compared to benchmarks, both in collision avoidance and reducing estimation covariance, affirming its robustness for complex robotic systems.

Keywords: Sensor fusion, Estimation
Abstract: Measurements from the global navigation satellite system (GNSS) and inertial measurement units (IMU) complement one another. Global position is observed from GNSS at low frequencies, and attitude and angular velocity are estimated from IMU data at high frequencies. However, in order to observe global attitude, there must be sufficient motion to excite each axis of the IMU, which may not be possible for large vehicles with constrained dynamics. We propose to model the trajectory estimate on the flat output space of a motion model whose heading is constrained to be in the direction of motion. This mitigates the need to observe heading from the measurements. In this method we use a continuous-time spline and optimize the control points such that the flat output trajectory fits the available GNSS and IMU measurements. We validate the proposed method with simulated data and show that it achieves a higher accuracy and lower solve time than continuous-time estimation on the configuration manifold SE(3).

Keywords: Identification
Abstract: This paper attacks practical difficulties in estimating parameters of a matrix fraction described system. Results on the sloppiness of a linear functional transformation (LFT) described descriptor system have been extended. Besides the assumptions that both the numerator and denominator polynomial matrices depend affinely on system parameters, and the denominator polynomial matrix is invertible, not any other hypothesis is adopted. An ellipsoid approximation is obtained for the parameter set constituted from all the parameter values, with the associated system frequency responses deviating within a specified range from those of the system with a known parameter value. For some significant application cases, explicit expressions were derived for the absolute and relative sloppiness. Comparisons with the well-known Fisher information matrix have also been performed through a mechanical system, revealing significant differences between them in quantifying parameter estimation hardness.

Keywords: Machine learning, Modeling
Abstract: The representation of a nonlinear dynamical system as infinite-dimensional linear operator over a Hilbert space of functions enables the study of the nonlinear system via pseudo-spectral analysis of the corresponding operator. In this paper, we develop a novel operator representation of discrete-time, control-affine nonlinear dynamical systems. We also demonstrate that this representation can be used to predict the behavior of the closed-loop system in response to a given feedback law. The representation is learned using recorded snapshots of the system state resulting from arbitrary, potentially open-loop control inputs. We thereby extend the predictive capabilities of dynamic mode decomposition to discrete-time nonlinear systems that are affine in control. We validate the method using two numerical experiments by predicting the response of a controlled Duffing oscillator to a known feedback law. The advantages of the developed method relative to existing techniques in the literature are also demonstrated.

Keywords: Machine learning, Biomedical, Biotechnology
Abstract: Due to its non-invasive nature, the Electromyography (EMG) signal has become crucial in predicting human motor intention and facilitating human-robot collaboration across various fields, including rehabilitation, assistive technologies, ergonomics, clinical diagnosis, and sport science. This study aims to enhance joint angle prediction using a two-step algorithm, leveraging a substantial dataset of twenty human subjects. The EMG data were acquired from four upper-limb muscles during an elbow flexion-extension and shoulder abduction-adduction task using Kinect V2 and Noraxon’s TeleMyo Mini DTS system. In stage one, a Support Vector Machine (SVM) is used to predict joint angles directly, followed by refining predictions with a Long Short-Term Memory (LSTM) network. The impact of different sliding window sizes on algorithm performance is analyzed, and a statistical comparison between direct angle prediction and the two-stage approach is conducted to assess the proposed method’s effectiveness. The proposed SVM-LSTM algorithm consistently yields enhanced Root Mean Square Error (RMSE) values (ranging from 8.45o to 9.36o) for both elbow and shoulder angles, compared to RMSE values reported in the literature (ranging from 8.2o to 19.04o). Statistical analysis through Friedman Analysis of Variance (ANOVA) underscores significant differences between SVM-direct and SVM-LSTM for both elbow and shoulder angles (p < 0.001). In the dynamic field of EMG-based joint angle prediction, these results hold profound significance for applications like prosthetics and rehabilitation robotics, where precise joint angle estimations are crucial.

Keywords: Machine learning, Chaotic systems, Reduced order modeling
Abstract: This paper examines the use of operator-theoretic approaches to the analysis of chaotic systems through the lens of their unstable periodic orbits (UPOs). Our approach involves three data-driven steps for detecting, identifying, and stabilizing UPOs. We demonstrate the use of kernel integral operators within delay coordinates as an innovative method for UPO detection. For identifying the dynamic behavior associated with each individual UPO, we utilize the Koopman operator to present the dynamics as linear equations in the space of Koopman eigenfunctions. This allows for characterizing the chaotic attractor by investigating its principal dynamical modes across varying UPOs. We extend this methodology into an interpretable machine learning framework aimed at stabilizing strange attractors on their UPOs. To illustrate the efficacy of our approach, we apply it to the Lorenz attractor as a case study.

Keywords: Machine learning, Formal verification/synthesis, Agents-based systems
Abstract: In causal reinforcement learning (RL), counterfactual reasoning deals with “what if” situations and allows for investigating the potential consequences of actions or events that did not actually happen. In this paper, we combine counterfactual reasoning and reinforcement learning (RL) and propose Counterfactually-Guided Causal Reinforcement Learning with Reward Machines (CGC-RL). In CGC-RL, using observational data, we first compute the optimal textit{counterfactual sequence} with the highest probability of completing a given task. Then, we construct an RM compatible with the textit{counterfactual sequence}. We use the constructed RM to apply dynamic potential-based reward shaping to encourage the agent to follow the textit{counterfactual sequence}. We prove the policy-invariance under dynamic reward shaping with RMs. Finally, we implement CGC-RL in one case study and compare the results with three baselines. Our results show that CGC-RL outperforms the baselines.

Keywords: Machine learning, Lyapunov methods, Chemical process control
Abstract: Machine learning (ML) methods such as neural networks provide an efficient way to build nonlinear dynamic models from data that can be used in the model predictive control system. In this work, we develop – for the first time – the method of ML modeling of nonlinear processes with Lyapunov stability guarantees, and the theory of its generalization performance and closed-loop stability. By designing a novel loss function that accounts for Lyapunov stability conditions in the training phase, the resulting Lyapunov-stable neural network (LSNN) is able to capture the nonlinear dynamics and guarantee closed-loop stability when incorporated in a stabilizing controller simultaneously. Provable closed-loop stability properties are derived based on the analysis of its generalization performance using statistical learning theory. Finally, a chemical reactor example is used to demonstrate the efficacy of the proposed ML modeling method.

Keywords: Human-in-the-loop control, Stochastic optimal control, Machine learning
Abstract: In many real-world scenarios involving high-stakes and safety implications, a human decision-maker (HDM) may receive recommendations from an artificial intelligence while holding the ultimate responsibility of making decisions. In this letter, we develop an "adherence-aware Q-learning" algorithm to address this problem. The algorithm learns the "adherence level" that captures the frequency with which an HDM follows the recommended actions and derives the best recommendation policy in real time. We prove the convergence of the proposed Q-learning algorithm to the optimal value and evaluate its performance across various scenarios.

Keywords: Networked control systems, Distributed control
Abstract: Although event-triggered architectures reduce the total number of agent-to-agent information exchange in distributed control of multiagent systems, their closed-loop performance can significantly deviate from their ideal, non-event-triggered counterparts. To address this issue, this paper presents a decentralized corrective signal, which is system-theoretically designed to achieve the ideal closed-loop multiagent system performance in event-triggered distributed control. In particular, the stability of the event-triggered closed-loop multiagent system with this decentralized corrective signal is first given and it is then demonstrated that the closed-loop performance approaches its ideal one as the gain of the decentralized corrective signal increases. A numerical example is also presented to illustrate the effectiveness of our approach when it is applied to a team of nonholonomic agents. While the results of this paper focus on performance recovery in distributed control of leader-follower multiagent systems over connected and undirected graphs, the presented decentralized corrective signal can be adapted for use in other distributed control architectures (e.g., consensus, containment, and formation control) as well as for other graph topologies, with minor changes.

Keywords: Networked control systems, Agents-based systems
Abstract: This paper solves the problem of estimating a single target position by a group of agents in two-dimensional space. Each agent is assumed to know its own position and to measure only the bearing to the target. We develop a distributed estimation scheme that processes individual bearing measurements by utilizing continuous-time dynamics with communication among the agents. Under the assumption of connectivity on the communication graph, it is shown that, even if each bearing measurement does not satisfy the persistency of excitation condition, each agent is able to estimate the target position provided that the agents are properly positioned. We examine this condition in terms of spatial excitation and its relationship with geometric configuration of the agents. Additionally, we analyze the convergence rate of the estimation error. Simulation results are provided to illustrate the performance of the proposed estimator.

Keywords: Networked control systems, Control of networks, Network analysis and control
Abstract: This paper explores the controllability of linear descriptor systems on graphs. These systems, which are a generalization of linear time-invariant systems, are widely used for modeling various real-life phenomena with equations that include both algebraic and differential terms. There are three main types of controllability for these systems: controllability within a reachable set (R-controllability), complete controllability (C-controllability), and impulse controllability (I-controllability). This work focuses on strong structural controllability, which refers to the ability to control a family of linear descriptor systems with the same underlying directed graph structure. To this end, different graphs are defined, and a correspondence between zero forcing sets of these graphs and the sets of control nodes required for strong structural R-controllability, C-controllability, and I-controllability is established.

Keywords: Cooperative control, Machine learning, Networked control systems
Abstract: This work presents an innovative learning-based approach to tackle the tracking control problem of Euler-Lagrange multi-agent systems with partially unknown dynamics operating under switching communication topologies. The approach leverages a correlation-aware cooperative algorithm framework built upon Gaussian process regression, which adeptly captures inter-agent correlations for uncertainty predictions. A standout feature is its exceptional efficiency in deriving the aggregation weights achieved by circumventing the computationally intensive posterior variance calculations. Through Lyapunov stability analysis, the distributed control law ensures bounded tracking errors with high probability. Simulation experiments validate the protocol's efficacy in effectively managing complex scenarios, establishing it as a promising solution for robust tracking control in multi-agent systems characterized by uncertain dynamics and dynamic communication structures.

Keywords: Networked control systems, Distributed control, Cooperative control
Abstract: This letter deals with a cloud-mediated self-triggered synchronization of physically coupled linear agents. In such a multi-agent system, each agent has not only information communication but also physical couplings with its neighboring agents. In the cloud-mediated control, each agent asynchronously communicates with its neighbors through a cloud repository. At every access time, each agent estimates current and future behaviors of its neighbors as well as of itself, and locally determines its next access time to the cloud according to a certain access rule. A major difficulty in the cloud-mediated control of physically coupled agents is that each agent cannot accurately estimate its neighbors’ behaviors because they are continuously influenced by other agents through the physical couplings. To overcome this difficulty, we propose a novel self-triggered synchronizing control law for the physically coupled linear agents. The proposed control law achieves bounded synchronization by appropriately bounding the adverse effects of the physical couplings.

Keywords: Autonomous robots, Autonomous systems, Robotics
Abstract: Safety assurance is critical in the planning and control of robotic systems. For robots operating in the real world, the safety-critical design often needs to explicitly address uncertainties and the pre-computed guarantees often rely on the assumption of the particular distribution of the uncertainty. However, it is difficult to characterize the actual uncertainty distribution beforehand and thus the established safety guarantee may be violated due to possible distribution mismatch. In this paper, we propose a novel safe control framework that provides a high-probability safety guarantee for stochastic dynamical systems following unknown distributions of motion noise. Specifically, this framework adopts adaptive conformal prediction to dynamically quantify the prediction uncertainty from online observations and combines that with the probabilistic extension of the control barrier functions to characterize the uncertainty-aware control constraints. By integrating the constraints in the model predictive control scheme, it allows robots to adaptively capture the true prediction uncertainty online in a distribution-free setting and enjoys formally provable high-probability safety assurance. Simulation results on multi-robot systems with stochastic single-integrator dynamics and unicycle dynamics are provided to demonstrate the effectiveness of our framework.

Keywords: Autonomous robots, Autonomous systems, Stability of nonlinear systems
Abstract: This paper presents a nonlinear control design for highly underactuated balance robots, where the number of unactuated degree-of-freedom is greater than that of actuated one. To address the challenge of simultaneously trajectory tracking and balancing, the control converts a robot dynamics into a series of cascaded subsystems and each of them is considered virtually actuated. We sequentially design and update the virtual and actual control inputs to incorporate the balance task such that the unactuated coordinates are balanced to their instantaneous equilibrium. The closed-loop dynamics are shown to be stable and the tracking errors exponentially converge towards a neighborhood near the origin. The simulation results demonstrate the effectiveness of the proposed control design by using a triple-inverted pendulum cart system.

Keywords: Autonomous robots, Constrained control, Multivehicle systems
Abstract: We consider the trajectory generation for a UGV and multiple UAVs that are tasked with monitoring certain specified target areas, where the former carries and re-charges the latter ones. We decouple the motion planning of UGVs and UAVs using reachable sets constructed from Lyapunov functions. The reachable sets are used as constraints for UGV trajectory generation resulting in existence guarantees of feasible UAVs trajectories with respect to flight and energy constraints. The reachable sets also provide an initial trajectory for UAVs rendezvous from and launch to the targets, which may be refined by optimization. We show simulation results of a case study with multiple UAVs monitoring multiple target sites.

Keywords: Autonomous robots, Control applications, Stochastic optimal control
Abstract: Target tracking has numerous significant civilian and military applications, and maintaining the visibility of the target plays a vital role in ensuring the success of the tracking task. Existing visibility-aware planners primarily focus on keeping the target within the limited field of view of an onboard sensor and avoiding obstacle occlusion. However, the negative impact of system uncertainty is often neglected, rendering the planners delicate to uncertainties in practice. To bridge the gap, this work proposes a model predictive control (MPC)-based trajectory planner for real-time visibility-aware and safe target tracking in the presence of system uncertainty. For more accurate target motion prediction, we introduce the concept of belief-space probability of detection (BPOD) to measure the predictive visibility of the target under stochastic robot and target states. An Extended Kalman Filter variant incorporating BPOD is developed to predict target belief state under uncertain visibility within the planning horizon. To reach real-time trajectory planning, we propose a computationally efficient algorithm to uniformly calculate both BPOD and the chance-constrained collision risk by utilizing linearized signed distance function (SDF), and subsequently solve the MPC problem by sequential convex programming. Extensive simulation results with benchmark comparisons show the capacity of the proposed approach to robustly maintain the visibility of the target under high system uncertainty. The practicality of the proposed trajectory planner is validated by real-world experiments.

Keywords: Autonomous robots, Markov processes, Intelligent systems
Abstract: Motivated by agility, 3D mobility, and low-risk operation compared to human-operated management systems of autonomous unmanned aerial vehicles (UAVs), this work studies UAV-based active wildfire monitoring where a UAV detects fire incidents in remote areas and tracks the fire frontline. A UAV path planning solution is proposed considering realistic wildfire management missions, where a single low-altitude drone with limited power and flight time is available. Noting the limited field of view of commercial low-altitude UAVs, the problem formulates as a partially observable Markov decision process (POMDP), in which wildfire progression outside the field of view causes inaccurate state representation that prevents the UAV from finding the optimal path to track the fire front in limited time. Common deep reinforcement learning (DRL)-based trajectory planning solutions require diverse drone-recorded wildfire data to generalize pre-trained models to real-time systems, which is not currently available at a diverse and standard scale. To narrow down the gap caused by partial observability in the space of possible policies, a belief-based state representation with broad, extensive simulated data is proposed where the beliefs (i.e., ignition probabilities of different grid areas) are updated using a Bayesian framework for the cells within the field of view. The performance of the proposed solution in terms of the ratio of detected fire cells and monitored ignited area (MIA) is evaluated in a complex fire scenario with multiple rapidly growing fire batches, indicating that the belief state representation outperforms the observation state representation both in fire coverage and the distance to fire frontline.

Keywords: Autonomous robots, Multivehicle systems, Constrained control
Abstract: This paper proposes a control solution to achieve collision-free platooning control of input-constrained mobile robots. The platooning policy is based on a leader-follower approach where the leader tracks a reference trajectory while followers track the leader's pose with an inter-agent delay. First, the leader and the follower kinematic models are feedback linearized and the platoon's error dynamics and input constraints characterized. Then, a set-theoretic model predictive control strategy is proposed to address the platooning trajectory tracking control problem. An ad-hoc collision avoidance policy is also proposed to guarantee collision avoidance amongst the agents. Finally, the effectiveness of the proposed control architecture is validated through experiments performed on a formation of Khepera IV differential drive robots.

Keywords: Estimation, Autonomous vehicles, Switched systems
Abstract: A new interval observer (IO) for uncertain switched linear dynamical systems is derived in this paper. We show that estimation accuracy and robustness can be achieved by combining Input-to-State Stability (ISS) concept and common quadratic Lyapunov function. The stability and positivity conditions are expressed in terms of Linear Matrix Inequalities (LMIs). This new observer overcomes the drawbacks of existing IO's, which are designed based on similarity transformations. The strengths of the proposed observer structure lie in having additional decision variables. The design procedure is very simple and uniform, it does not require changes of coordinates, which are costly in terms of computational time. The increased robustness of the presented method, compared to other commonly used approaches, is demonstrated by accurately estimating, at each time instant, the interval including vehicle sideslip angle and yaw rate. The proposed algorithm shows high potential for practical implementation.

Keywords: Estimation, Filtering, Stochastic systems
Abstract: In estimation, control, and machine learning under uncertainties, latent variables are usually described by a probability density function (pdf). The optimal reconstruction of a continuous pdf from given samples or moments is an important and ubiquitous task. Unfortunately, it typically results in an underdetermined optimization problem, as the pdf is not fully constrained by the given samples or moments. For regularization, we use Fisher Information (FI) that acts as a roughness measure, i.e., selects the smoothest pdf fulfilling the constraints, in an information-theoretic sense. For the important class of mixture densities, FI can only be computed numerically. In this paper, we derive a closed-form solution for FI for mixtures by transforming the problem to the space cal R of root mixtures (RMs). This results in a tandem processing scheme simultaneously working in the original mixture space cal M and the corresponding RM space: The density parameters are optimized in root mixture space based on the closed-form FI. The desired constraints are evaluated in the original mixture space cal M.

Keywords: Estimation, Identification, LMIs
Abstract: We study the problem of identifying a linear time-varying output map from measurements and linear time-varying system states, which are perturbed with Gaussian observation noise and process uncertainty, respectively. Employing a stochastic model as prior knowledge for the parameters of the unknown output map, we reconstruct their estimates from input/output pairs via a Bayesian approach to optimize the posterior probability density of the output map parameters. The resulting problem is a non-convex optimization, for which we propose a tractable linear matrix inequalities approximation to warm-start a first-order subsequent method. The efficacy of our algorithm is shown experimentally against classical Expectation Maximization and Dual Kalman Smoother approaches.

Keywords: Estimation, Kalman filtering, Sensor fusion
Abstract: In many applications, attitude estimation algorithms rely mainly on magnetic and inertial measurements from MARG sensors (consisting of a magnetometer, a gyroscope, and an accelerometer). One of the main challenges facing these algorithms is that the accelerometer measures both gravity and an unknown external acceleration, while these algorithms assume that the accelerometer measures only the gravity. In this paper, an attitude estimation algorithm on the special orthogonal group SO(3) is designed, considering the external acceleration as an unknown input with direct feedthrough to the output, with a local approximation approach. The proposed algorithm is validated through Monte Carlo simulations and real datasets, demonstrating better accuracy and enhanced performance than existing solutions.

Keywords: Estimation, Lyapunov methods
Abstract: We formulate the problem of fake news detection using distributed fact-checkers (agents) with unknown reliability. The stream of news/statements is modeled as an independent and identically distributed binary source (to represent true and false statements). Upon observing a news, agent i labels the news as true or false that reflects the true validity of the statement with some probability 1-pi_i. In other words, agent i misclassified each statement with error probability pi_iin (0,1), where the parameter pi_i models the (un)trustworthiness of agent i. We present an algorithm to learn the unreliability parameters, resulting in a distributed fact checking algorithm. For the two-agent case, we extensively analyze the discrete-time limit of our algorithm.

Keywords: Estimation, Machine learning, Chemical process control
Abstract: Data-driven predictive models for end-point quality variables are important tools in industrial process modeling. However, establishing an effective predictive model with limited labeled data remains challenging. Transfer learning (TL) offers a solution by leveraging knowledge from similar but different tasks. This paper introduces a novel TL-based predictive model, domain-adaptation parallel stacked autoencoders (DA-PSAE), which can extract and accumulate knowledge from multiple similar processes. First, a parallel model structure is designed for the simultaneous extraction of static and plastic latent features. Besides, an effective TL-based training strategy is proposed, which utilizes data from multiple similar processes. The proposed model is applied to a sulfur recovery unit composed of four parallel sub-units. Experimental results verify the effectiveness of the proposed model.

Keywords: Optimization algorithms, Linear systems, Robust control
Abstract: We study the effect of averaging on convergence speed and noise amplification of the heavy-ball algorithm in the presence of additive white stochastic disturbances. For strongly convex quadratic problems, we show that averaging over the entire algorithmic history eliminates steady-state variance of the averaged output at the expense of slowing down convergence to a sub-linear rate. In contrast, finite window averaging converges with a linear rate but it leads to a non-zero value of the steady-state variance. While this value is smaller than the steady-state variance of the iterates of the heavy-ball algorithm, it has the same orderwise dependence on the condition number. We also show that the finite window averaging increases the upper bound on the expected error at iteration t by a constant factor that depends on the length of the averaging window.

Keywords: Optimization algorithms, Machine learning, Optimal control
Abstract: Online learning algorithms for dynamical systems provide finite time guarantees for control in the presence of sequentially revealed cost functions. We pose the classical linear quadratic tracking problem in the framework of online optimization where the time-varying reference state is unknown a priori and is revealed after the applied control input. We show the equivalence of this problem to the control of linear systems subject to adversarial disturbances and propose a novel online gradient descent-based algorithm to achieve efficient tracking in finite time. We provide a dynamic regret upper bound scaling linearly with the path length of the reference trajectory and a numerical example to corroborate the theoretical guarantees.

Keywords: Optimization algorithms, Optimization, Lyapunov methods
Abstract: We study the geometric convergence rate of discrete-time primal-dual algorithms for solving strongly-convex equality-constrained optimization problems. Our approach separates the primal-dual algorithm into an interconnection of two exponentially stable systems, and a composite Lyapunov approach is used to establish stability of the interconnection and provide new bounds on the geometric rate of convergence. Analogous convergence results are developed for two variations of the primal-dual algorithm: an extrapolated version which accelerates convergence, and an inner-loop version which interpolates between the vanilla primal-dual method and dual ascent. The obtained bounds are compared and contrasted with existing bounds from the literature.

Keywords: Optimization algorithms, Network analysis and control, Machine learning
Abstract: This paper addresses the problem of nonconvex nonsmooth decentralised optimisation in multi-agent networks with undirected connected communication graphs. Our contribution lies in introducing an algorithmic framework designed for the distributed minimisation of the sum of a smooth (possibly nonconvex and non-separable) function and a convex (possibly nonsmooth and non-separable) regulariser. The proposed algorithm can be seen as a modified version of the ADMM algorithm where, at each step, an ``inner loop'' needs to be iterated for a number of iterations. The role of the inner loop is to aggregate and disseminate information across the network. We observe that a naive decentralised approach (one iteration of the inner loop) may not converge. We establish the asymptotic convergence of the proposed algorithm to the set of stationary points of the nonconvex problem where the number of iterations of the inner loop increases logarithmically with the step count of the ADMM algorithm. We present numerical results demonstrating the proposed method's correctness and performance.

Keywords: Optimization algorithms, Stochastic optimal control, Control of networks
Abstract: We present a method to transform any optimal stopping time problem with an underlying tree structure into an s-t min-cut problem on the same tree but with modified capacities, the details of which are lacking in existing optimal stopping time research. We also show that any s-t min/max-cut problem on a tree has an equivalent optimal stopping problem formulation. We provide a dynamic programming algorithm to solve this problem and also perform a sensitivity analysis on it. Our results imply that the s-t max-cut problem on a tree can be solved using an algorithm with runtime that is linear in the tree size.

Keywords: Optimization algorithms, Time-varying systems, Distributed control
Abstract: Resource allocation problems are ubiquitous across different engineering applications where a fixed and limited asset needs to be optimally allocated among multiple agents in an online manner. While a plethora of algorithms exist for the solution of such problems, most existing approaches are restricted to either centralized initializations or the use of multiple auxiliary states that increase the complexity of the dynamics. In this paper, we tackle these challenges by introducing a novel and simple approach for the solution of decentralized resource allocation problems in multi-agent networked systems. The approach relies on a family of time-varying Laplacian gradient flows with vanishing anchors that remove any restriction on the initialization of the dynamics, without the need of incorporating additional auxiliary states. The convergence properties of the proposed dynamics are studied for a general class of time-varying vanishing gains. Analytical and numerical examples are also presented.

Keywords: Energy systems, Estimation, Optimal control
Abstract: Optimal power management of battery energy storage systems (BESS) is crucial for their safe and efficient operation. Numerical optimization techniques are frequently utilized to solve the optimal power management problems. However, these techniques often fall short of delivering real-time solutions for large-scale BESS due to their computational complexity. To address this issue, this paper proposes a computationally efficient approach. We introduce a new set of decision variables called power-sharing ratios corresponding to each cell, indicating their allocated power share from the output power demand. We then formulate an optimal power management problem to minimize the system-wide power losses while ensuring compliance with safety, balancing, and power supply-demand match constraints. To efficiently solve this problem, a parametrized control policy is designed and leveraged to transform the optimal power management problem into a parameter estimation problem. We then implement the ensemble Kalman inversion to estimate the optimal parameter set. The proposed approach significantly reduces computational requirements due to 1) the much lower dimensionality of the decision parameters and 2) the estimation treatment of the optimal power management problem. Finally, we conduct extensive simulations to validate the effectiveness of the proposed approach. The results show promise in accuracy and computation time compared with explored numerical optimization techniques.

Keywords: Optimal control, Energy systems, Optimization
Abstract: Electric vertical takeoff and landing (eVTOL) aircraft operation often requires the battery packs to be charged and thermally conditioned to a given range of state-of-charge (SOC) and temperature for the next flight in minimum time to maximize aircraft utility. Where charging favors higher temperatures than what is required at launch, an optimal solution must trade-off faster charging times with the required thermal conditioning time. Here we develop a coupled electro-thermal battery, charger, and cooling system model, formulate an optimal control problem for fast charging, and solve it using a direct collocation approach. The resulting algorithm simultaneously provides the optimal current and thermal conditioning profiles that bring the pack from a starting temperature and SOC to their respective targets in minimum time, while ensuring degradation and safety constraints are satisfied. Insights from the results can be used to develop simple on-line control algorithms for charging with active thermal management that are critical for eVTOL applications.

Keywords: Energy systems, Optimization
Abstract: Dispatch of a grid energy storage system for arbitrage is typically formulated into a rolling-horizon optimization problem that includes a battery aging model within the cost function. Quantifying degradation as a depreciation cost in the objective can increase overall profits by extending lifetime. However, depreciation is just a proxy metric for battery aging; it is used because simulating the entire system life is challenging due to computational complexity and the absence of decades of future data. In cases where the depreciation cost does not match the loss of possible future revenue, different optimal usage profiles result and this reduces overall profit significantly compared to the best case (e.g., by 30–50%). Representing battery degradation perfectly within the rolling-horizon optimization does not resolve this—in addition, the economic cost of degradation throughout life should be carefully considered. For energy arbitrage, optimal economic dispatch requires a trade-off between overuse, leading to high return rate but short lifetime, vs. underuse, leading to a long but not profitable life. We reveal the intuition behind selecting representative costs for the objective function, and propose a simple moving average filter method to estimate degradation cost. Results show that this better captures peak revenue, assuming reliable price forecasts are available.

Keywords: Optimization, Energy systems, Transportation networks
Abstract: Decarbonization and electrification of long-haul trucks are notoriously difficult due to the high energy demand and limited gravimetric energy density of lithium-ion cells. In this study, we investigate the optimal deployment and operation of a grid-connected battery swapping station (BSS) for electric long-haul trucks as a mixed-integer optimization problem. We construct a model for reliably meeting customer energy needs while providing grid services, to demonstrate the business case and the operation of such a system. The impact of optimal sizing of the station is explored. A comparative study of two alternative charging infrastructure solutions using battery swapping and fast charging stations (FCS) has been performed through numerical experiments. We examine ton-mile-per-hour as a metric for the relative efficiency of cargo movement and the labor requirements. We find that FCS works better for large battery packs with short and regional hauls, and BSS is well suited for relatively medium-sized battery packs and long-haul truck movements. The study also found that the potential to use smaller batteries, and the expansion of the network can increase the battery utilization efficiency of BSS, which may be able to approach a similar level to the utilization efficiency of FCS.

Keywords: Estimation, Model Validation, Energy systems
Abstract: Large battery systems include parallel-connected cells and modules, and these can exhibit complex and unexpected behaviours. In this paper, we investigate parallel-connected battery equivalent circuit models and show that cell inhomogeneity leads to differences between parallel groups that are not typically accounted for in battery management system (BMS) models. The theoretical analysis is complemented by experimental results using impedance data to parameterise models from individual cells, and from a pack assembled from the individual cells. We show that cell-level parameterisation is sufficient for parameterising a pack model---but only if parasitic impedances and inhomogeneity between cells is well understood.

Keywords: Automotive control, Control applications, Energy systems
Abstract: This study proposes an energy-critical control scheme for road vehicles. The framework is designed to ensure an energy-critical goal that is defined in the form of an upper bound on the energy consumption. In order to respond to the changing traffic environment, this bound is changed based on the preceding vehicle's energy consumption. Control barrier functions are used to obtain a condition on the control input such that the energy-critical goal is formally guaranteed. This condition is used in a quadratic program scheme as a constraint, and the resulting controller intervenes the driver input in a minimally-invasive fashion. The effect of the energy-critical controller is demonstrated in data-based simulations, where up to 25% improvement in energy consumption is observed compared to an energy-inefficient driver, without drastically changing the car-following performance.

Keywords: Agents-based systems, Machine learning, Constrained control
Abstract: This research aims to develop a new learning-based flocking control framework that ensures inter-agent free collision. To achieve this goal, a leader-following flocking control based on a deep Q-network (DQN) is designed to comply with the three Reynolds’ flocking rules. However, due to the inherent conflict between the navigation attraction and inter agent repulsion in the leader-following flocking scenario, there exists a potential risk of inter-agent collisions, particularly with limited training episodes. Failure to prevent such collision not only caused penalties in training but could lead to damage when the proposed control framework is executed on hardware. To address this issue, a control barrier function (CBF) is incorporated into the learning strategy to ensure collision-free flocking behavior. Moreover, the proposed learning framework with CBF enhances training efficiency and reduces the complexity of reward function design and tuning. Simulation results demonstrate the effectiveness and benefits of the proposed learning methodology and control framework.

Keywords: Adaptive control, Automotive control, Direct adaptive control
Abstract: Existing adaptive control designs for vehicle steering-by-wire (SbW) systems mainly rely on quadratic Lyapunov functions, providing (global) stability and, at best, (global) asymptotic convergence of certain closed-loop signals. However, these approaches generally lack assurance in transient performance. In this paper, we introduce a novel adaptive control scheme aimed to enhance and guarantee the transient performance of the adaptive SbW control system. This approach integrates a varying-degree Lyapunov function with deterministic robust control. The new adaptive control scheme is derived in a general context, applicable to a class of single-input, parametrically uncertain, nonlinear dynamic systems in Brunovsky form. We then apply this general theoretical result to develop an adaptive controller for the SbW system. Using a high-fidelity moving-base driving simulator, we demonstrate the transient performance improvement of the new adaptive SbW controller compared to a baseline method.

Keywords: Automotive control, Estimation, Mechatronics
Abstract: Teleoperation of a vehicle requires displaying the road environment of the remote vehicle accurately on a teleoperation station, so that a human teleoperator can use the display to control the vehicle safely and efficiently. Limited bandwidth and latencies in wireless communication may prevent transmission of camera images and Lidar data at a sufficiently high frequency for rapid updates of the display. This paper describes how frequent transmission of just GPS and IMU data can enable accurate vehicle position and orientation estimation with which realistic intermediate updates of the remote vehicle environment can be provided. A nonlinear dynamic motion model and an extended Kalman filter are utilized for estimating vehicle position and orientation. A study with 5 human subjects is used to compare steering control of a remote vehicle with and without intermediate position updates. Experimental data show that a 0.5 second delay in real-time display makes it extremely difficult for a human teleoperator to control the vehicle to stay in its lane on curved roads. However, using an estimation-based predictive display system to update the vehicle position and orientation with respect to the road environment enables safe remote vehicle control with almost as accurate a performance as the delay-free case.

Keywords: Traffic control, Cooperative control, Autonomous systems
Abstract: Connected and automated vehicles (CAVs) have been widely applied to vehicle-based traffic control in mixed-autonomy systems that consist of CAVs and human-driven vehicles (HVs). The control designs of CAVs allow them to drive smoothly, cooperate, and stabilize the flow of traffic. However, these controllers must also ensure safe behaviors for CAVs as well as consider their potential impact on the safety of following HVs. In this paper, we propose nonlinear controllers for a pair of CAVs that respond to each other whilst traveling amongst HVs. The controllers seek to stabilize traffic, while the safety of CAVs in terms of~front-end collisions is formally guaranteed via control barrier functions. We analyze how the coordination of CAVs affects stability and safety in the mixed traffic system. The efficacy of the proposed controllers is demonstrated by numerical simulations.

Keywords: Energy systems, Machine learning, Modeling
Abstract: Batteries are dynamic systems with complicated nonlinear aging, highly dependent on cell design, chemistry, manufacturing, and operational conditions. Prediction of battery cycle life and estimation of aging states is important to accelerate battery R&D, testing, and to further the understanding of how batteries degrade. Beyond testing, battery management systems rely on real-time models and onboard diagnostics and prognostics for safe operation. Estimating the state of health and remaining useful life of a battery is important to optimize performance and use resources optimally.

This tutorial begins with an overview of first-principles, machine learning, and hybrid battery models. Then, a typical pipeline for the development of interpretable, machine learning models is explained and showcased for cycle life prediction from laboratory testing data. We highlight the challenges of machine learning models, motivating the incorporation of physics in hybrid modeling approaches, which are needed to decipher the aging trajectory of batteries but require more data and further work on the physics of battery degradation. The tutorial closes with a discussion on generalization and further research directions.

Keywords: Fault diagnosis, Machine learning, Fault detection
Abstract: This article provides a tutorial on data-driven methods for diagnostics and prognostics. The methods span the range from classical methods to machine learning-based methods including ensemble methods and support vector machine. The main challenges associated with diagnostics/prognostics are discussed, and decision rules are provided for selecting which data-driven modeling method to apply to a particular dataset based on the characteristics of the data. The autonomous selection of data-driven methods is described, as an application-specific automated machine learning (AutoML) approach. A case study of a simulated chemical plant provided by Tennessee Eastman Company is used to illustrate the characteristics of the data-driven methods for diagnostics and prognostics.

Keywords: Aerospace, Machine learning, Intelligent systems
Abstract: This paper presents a comparative analysis of different machine learning algorithms for aircraft loss of control (LOC) prediction. Aircraft LOC is a flight dynamics and control phenomenon that has been associated with a significant number of aircraft accidents worldwide. The NASA Generic Transport Model (GTM) is used as a representative aircraft to study the efficacy of machine learning based prediction algorithms under LOC scenarios. Extensive simulation of the GTM model in normal and LOC flight conditions are carried out to generate training and testing data for the machine learning algorithms. Different machine learning methods, viz. neural networks, support vector machines, decision trees, ensemble subspace kNN, ensemble boosted trees, and ensemble bagged trees are examined for their effectiveness in predicting LOC. A multi class classification problem is solved to build the machine learning models. It is observed from the analysis of prediction results that neural network method provides the best predictive model. Furthermore, the neural network predictive model performs quite well in correctly classifying the flight conditions under LOC, with an overall accuracy of 97 %

Keywords: Aerospace, Optimal control, Optimization algorithms
Abstract: Most of the optimal guidance problems can be formulated as nonconvex optimization problems, which can be solved indirectly by relaxation, convexification, or linearization. Although these methods are guaranteed to converge to the global optimum of the modified problems, the obtained solution may not guarantee global optimality or even the feasibility of the original nonconvex problems. In this paper, we propose a computational optimal guidance approach that directly handles the nonconvex constraints encountered in formulating the guidance problems. The proposed computational guidance approach alternately solves the least squares problems and projects the solution onto nonconvex feasible sets, which rapidly converges to feasible suboptimal solutions or sometimes to the globally optimal solutions. The proposed algorithm is verified via a series of numerical simulations on impact angle guidance problems under state dependent maneuver vector constraints, and it is demonstrated that the proposed algorithm provides superior guidance performance than conventional techniques.

Keywords: Aerospace, Robotics, Autonomous robots
Abstract: Some inspection missions for quadrotors require a previous definition of the trajectories before being tracked. The well design of the trajectory can contribute for the success of the mission when aggressive maneuvers are demanded. In this paper, we propose a solution for designing and tracking unconventional trajectories requiring aerobatics maneuvers in quadrotors. The trajectory is based on the Hopf bifurcations and the sliding mode approach to form spiral loops in 3D around a pillar that need to be inspected by a quadrotor. The control architecture is based on the sliding mode methodology adding new parameters for achieving a desired angular velocity around the stable limit cycle solution. The proposed solution is validated in simulations and numerical results illustrate the good performance of the proposed trajectory-tracking control scheme.

Keywords: Aerospace, Robotics, Hybrid systems
Abstract: Control barrier functions are commonly used for providing safety guarantees to a controlled dynamical system, such as a guarantee of collision avoidance. This work leverages the recent advancements in multiple higher-order control barrier functions to develop a control strategy for a free-flying space manipulator system (SMS) with an arbitrary number of redundant manipulator arms to capture a non-cooperative resident space object (RSO). The impact of this result is a real-time controller capable of being used on a computationally constrained system while satisfying several safety constraints necessary for the approach, synchronization, and capture of the RSO.

Keywords: Aerospace, Variable-structure/sliding-mode control
Abstract: This paper introduces a novel impact angle-constrained guidance law for tackling the problem of stationary target interception within a planar framework. The guidance law leverages a class of bounded curvature spirals as the fundamental geometric framework, allowing the pursuer to enforce a desired impact angle by choosing a suitable spiral trajectory. Another key aspect that this paper addresses is that of minimizing the maximum lateral acceleration required by the pursuer since this is a crucial determinant of a guidance law's effectiveness. The geometric law is implemented using a robust nonlinear super-twisting sliding-mode control-based guidance command to ensure finite-time interception of targets. Extensive numerical simulations have been performed to validate the theoretical results and the robustness of the guidance law.

Keywords: Aerospace
Abstract: In this letter, we propose a three-dimensional nonlinear guidance strategy that guarantees the interception of stationary targets at the pre-specified impact time while accounting for the onboard seeker’s field-of-view constraints. The proposed approach for designing guidance law does not require a time-to-go estimation. It utilizes range-to-go, free from time-to-go approximations, allowing the interceptor to control the time of target interception. The guidance command, derived using the prescribed constraint function and sliding mode control considering coupled and nonlinear engagement kinematics, continuously maintains the target within its field-of-view and achieves guidance objectives. The proposed guidance strategy performs effectively, even for large deviations from the collision course, where pitch and yaw channels may be strongly coupled. Finally, the efficacy of the proposed guidance strategy is demonstrated using extensive numerical simulations.

Keywords: Autonomous robots, Optimization algorithms, Stochastic systems
Abstract: Model Predictive Path Integral (MPPI) control has proven to be a powerful tool for the control of uncertain systems. One important limitation of the baseline MPPI algorithm is that it does not utilize simulated trajectories to their fullest extent. For one, it assumes that the average of all trajectories weighted by their performance index will be a safe trajectory. In this paper, multiple examples are shown where the previous assumption does not hold, and a trajectory clustering technique is presented that reduces the chances of the weighted average crossing in an unsafe region. Secondly, MPPI does not account for dynamic obstacles, so the authors put forward a novel cost function that accounts for dynamic obstacles without adding significant computation time to the overall algorithm. The novel contributions proposed in this paper were evaluated with extensive simulations to demonstrate improvements upon the state-of-the-art MPPI techniques.

Keywords: Aerospace, Control applications, Variable-structure/sliding-mode control
Abstract: In this paper, we address the problem of robust guidance for a three-body pursuit-evasion problem involving a target, a defender, and an attacker, where the goal is to safeguard the target (an aircraft) from the attacker, using the defender. Thus, the target and the defender may act as a team of vehicles whose common goal is to safeguard the target from the attacker. Additionally, it is desired to maintain necessary constraints on certain states. An adaptive guidance law is proposed for the target-defender team to ensure optimal performance within this set-up. The target-defender dynamics are considered for designing the guidance law with only approximate knowledge of the unknown parameters of the attacker. A robust guidance command is designed using a super-twisting algorithm with adaptive gains. The unknown disturbance bounds required for control implementation are obtained using a novel norm observer. We consider the complete nonlinear model of the system for control design, thereby enabling a wider domain of applicability of the proposed approach. The convergence is established using Lyapunov analysis, which also ensures satisfaction of the required state constraints. Relevant simulation results are presented to study the efficacy of the proposed approach.

Keywords: Optimal control, Optimization algorithms, Linear systems
Abstract: This article presents a method to automatically generate energy-optimal trajectories for systems with linear dynamics, linear constraints, and a quadratic cost functional (LQ systems). First, using recent advancements in optimal control, we derive the optimal motion primitive generator for LQ systems. This yields linear differential equations that describe all dynamical motion primitives that the optimal system trajectory follows. We demonstrate the performance of our approach in simulation on an energy-minimizing submersible robot with state and control constraints and compare our approach to an energy-optimizing Linear Quadratic Regulator.

Keywords: Human-in-the-loop control, Robotics, Machine learning
Abstract: In this paper, we develop a semi-autonomous full 3D robot operation system for a target reaching and holding task through Gaussian Process Regression (GPR). In order to mitigate the limited dimensionality of the human manipulable command, a virtual reality (VR) controller is employed as a command interface that links the operator’s own arm movements with those of the robot. We start by having a well-trained operator operate the robot from various initial positions away from an object to holding the object with grippers while directly looking at the manipulator, instead of through a 2D display. We then construct a GPR model of the operator, and design an automatic control law based on the mean function of the model, where we reveal the need for thinning out the training data to eliminate a computational issue stemming from excessive data. It is further pointed out that the high dimensionality of the robot motion poses a new challenge, namely a trade-off between data coverage and computational effort. To address this issue, we present a semi-autonomous system that blends manual control and automatic control based on the variance function of the GPR model. Finally, the proposed operation support system is demonstrated through experiments.

Keywords: Optimization, Cooperative control, Transportation networks
Abstract: Multi Agent Path Finding (MAPF) seeks the optimal set of paths for multiple agents from respective start to goal locations such that no paths conflict. We address the MAPF problem for a fleet of hybrid-fuel unmanned aerial vehicles which are subject to location-dependent noise restrictions. We solve this problem by searching a constraint tree for which the subproblem at each node is a set of shortest path problems subject to the noise and fuel constraints and conflict zone avoidance. A labeling algorithm is presented to solve this subproblem, including the conflict zones which are treated as dynamic obstacles. We present the experimental results of the algorithms for various graph sizes and number of agents.

Keywords: Cooperative control, Autonomous robots, Manufacturing systems
Abstract: This paper presents a novel fleet management strategy for battery-powered robot fleets tasked with intra-factory logistics in an autonomous manufacturing facility. In this environment, repetitive material handling operations are subject to real-world uncertainties such as blocked passages, and equipment or robot malfunctions. In such cases, centralized approaches enhance resilience by immediately adjusting the task allocation between the robots. To overcome the computational expense, a two-step methodology is proposed where the nominal problem is solved a priori using a Monte Carlo Tree Search algorithm for task allocation, resulting in a nominal search tree. When a disruption occurs, the nominal search tree is rapidly updated a posteriori with costs to the new problem while simultaneously generating feasible solutions. Computational experiments prove the real-time capability of the proposed algorithm for various scenarios and compare it with the case where the search tree is not used and the decentralized approach that does not attempt task reassignment.

Keywords: Game theory, Agents-based systems, Finance
Abstract: The condition when the Nash equilibrium set has a simple topological structure is presented. Specifically, we first define the terms `weakly simplicial' and `simplicial' for the Nash equilibrium set characterization problem (NESCP), which is a generalization of the terms for the Pareto optimality characterization problem. Then, we give a condition for when the NESCP is weakly simplicial so that the Nash equilibrium set is compact. After that, we present a condition for the NESCP to be simplicial, when each face of the simplices corresponds to a face of the Nash equilibrium set. Finally, we give a couple of examples to illustrate our results.

Keywords: Game theory, Agents-based systems
Abstract: In binary decision-makings, individuals often go for a common or rare action. In the framework of evolutionary game theory the best-response update rule can be used to model this dichotomy. Those who prefer a common action are called coordinators and those who prefer a rare one are called anticoordinators. A finite mixed population of the two types may undergo perpetual fluctuations, the characterization of which appears to be challenging. It is particularly unknown, whether the fluctuations scale with population size. To fill this gap, we approximate the discrete finite population dynamics of coordinators and anticoordinators with the corresponding mean dynamics in the form of (semi-)continuous differential inclusions. We show that the family of the state sequences of the discrete dynamics for increasing population sizes forms a generalized stochastic approximation process for the differential inclusion. On the other hand, we show that the differential inclusions always converge to an equilibrium. This implies that the reported perpetual fluctuations in the finite discrete dynamics of coordinators and anticoordinators do not scale as the population size do. The results encourage to first analyze the often simpler (semi-)continuous mean dynamics of discrete population dynamics as the continuous dynamics partly reveal the asymptotic behaviour of the discrete dynamics.

Keywords: Game theory, Agents-based systems
Abstract: In problems of multiagent coordination, it is known that equilibria arising from self-interested behavior can be highly inefficient. This inefficiency, quantified by the Price of Anarchy (PoA), has been studied for a wide range of problems. Typically, only worst-case boundsare known on the PoA, leaving significant questions: are highly inefficient equilibria likely? Are they stable? In this paper, we comprehensively prove that worst-case Nash equilibria in a large class of games are not stable, in a game-theoretically relevant sense: at these equilibria, every agent can switch to its system-optimal action without incurring any loss in individual payoff. We parameterize the inefficiency of equilibria by a new measure of stability which quantifies agents' aggregate "satisfaction" with their equilibrium choices; we show that if agents are highly "satisfied" with their actions at a particular equilibrium (i.e., the equilibrium is highly stable), then that equilibrium must be relatively efficient. Our proof techniques are highly general and apply to all lambda-mu-smooth games, which account for a large fraction of games with known PoA bounds including routing, scheduling, resource allocation, and auctions. We also perform numerical experiments to contextualize our work in specific games of interest.

Keywords: Game theory, Distributed control, Agents-based systems
Abstract: This paper investigates the Nash equilibrium seeking problem for aggregative games, where some players are under Byzantine attacks. Since the previous distributed algorithms might be seriously affected by Byzantine attacks that inject malicious information to the system, we propose a resilient distributed geometric-median based method to improve the robustness of the algorithm. When less than half of the players are Byzantine attackers and under strongly monotone assumption on the pseudo-gradient mapping, we show that the algorithm converges linearly to a neighborhood of the Nash Equilibrium with an error bound related to the ratio of Byzantine attackers. Finally, numerical examples are implemented to demonstrate the robustness of the proposed algorithm under various types of Byzantine attacks.

Keywords: Game theory, Distributed control, Stochastic optimal control
Abstract: This paper presents a novel distributed optimization algorithm for large-scale multi-agent systems (LS-MAS), particularly with a given fixed final density constraint. Although the Mean field game (MFG) theory provides a distribution solution to overcome the ``Curse of dimensionality" in LS-MAS, it significantly sacrifices LS-MAS optimality and also not be capable of achieving arbitrary fixed final probability density function (PDF) constraint. To overcome these challenges, a novel Imbalanced Mean-Field Game (Imb-MFG) theory is developed along with an adaptive PDF decomposition algorithm and distributed reinforcement learning. Specifically, an induction-based PDF parameter estimation is developed to decompose the final density constraints into multiple imbalanced norm distributions. Then, the Imb-MFG theory is designed by integrating multi-group MFG with a constrained K-means clustering algorithm. To solve the developed Imb-MFG and further obtain the distributed optimal solution, a multi-actor-critic-mass (Multi-ACM) algorithm is designed to learn the solution of multi-group coupled Hamilton-Jacobi-Bellman (HJB) and Fokker-Planck-Kolmogorov (FPK) equations simultaneously. Finally, the convergence of the developed Multi-ACM algorithm is guaranteed through Lyapunov analysis.

Keywords: Game theory, Hybrid systems, Optimal control
Abstract: In this paper, we formulate a two-player zero-sum game under dynamic constraints formulated in terms of a hybrid inclusion. The game consists of a min-max problem involving a cost functional associated to the actions and corresponding (potentially nonunique) solutions to the system. We present sufficient conditions given in terms of Hamilton-Jacobi- Isaacs-like equations to establish a bound on the worst-case cost under the optimal strategy and to exactly evaluate it. Under additional conditions, we show that the proposed optimal state-feedback laws render a set of interest pre-symptotically stable for the resulting hybrid closed-loop system. The results are illustrated in a numerical example.

Keywords: Automotive systems, Predictive control for nonlinear systems, Autonomous systems
Abstract: This paper presents a trajectory planning method for connected and automated vehicles (CAVs) operating on a two-lane urban road that includes signalized intersections with a mix of human-driven vehicles (HVs) and CAVs. The proposed approach aims to find the optimal trajectory for CAVs, considering factors such as energy consumption, travel time and passengers' comfort. To achieve this objective, each CAV utilizes data obtained from various sources. Vehicle-to-infrastructure (V2I) communication enables access to information such as traffic light cycle lengths, timings and positions. Vehicle-to-vehicle (V2V) communication allows CAVs to gather information from other CAVs, including their positions, velocities and predicted trajectories. Additionally, on-board sensors can also provide data about surrounding vehicles. Dynamic programming (DP) is employed to predict the velocity profile of the CAVs over a long horizon, incorporating information from traffic lights and the vehicle's specifications. The predicted velocity profile is then fed into a model predictive controller (MPC) to follow the CAV's trajectory. The MPC is capable of handling disturbances encountered during the maneuvers, such as stopping behind a red traffic light or performing lane changes. Extensive simulation studies demonstrate how various objectives in terms of saving in energy consumption and travel time are achieved.

Keywords: Identification for control, Mechatronics, Predictive control for nonlinear systems
Abstract: The topic of this paper is the design of a data-driven model predictive controller for a nonlinear ball-on-a-wheel system. A Koopman predictor is identified using simulation data of a ball-on-a-wheel system model. The challenge lies in using data, which is characterized by the instability of the given system. This limits the time horizon during which meaningful time series can be extracted. Within this scenario, the predictor is included in a model predictive control scheme. This controller successfully tracks a specified non-flat output trajectory subject to input and output constraints. Finally, the prediction quality of the Koopman predictor is compared with a linear predictor based on dynamic mode decomposition with control.

Keywords: Manufacturing systems and automation, Predictive control for nonlinear systems, Control applications
Abstract: This paper presents a modeling-control synthesis to address the quality control challenges in multistage manufacturing systems (MMSs). A new feedforward control scheme is developed to minimize the quality variations caused by process disturbances in MMSs. Notably, the control framework leverages a stochastic deep Koopman (SDK) model to capture the quality propagation mechanism in the MMSs, highlighted by its ability to transform the nonlinear propagation dynamics into a linear one. Two roll-to-roll case studies are presented to validate the proposed method and demonstrate its effectiveness. The overall method is suitable for nonlinear MMSs and does not require extensive expert knowledge.

Keywords: Neural networks, Predictive control for nonlinear systems, Network analysis and control
Abstract: It has been theoretically explained, through the notion of Neural Tangent Kernel, why the training error of overparameterized networks converges linearly to 0. In this work, we focus on the case of small (or underparameterized) networks. An advantage of small networks is that they are faster to train while often retaining sufficient precision to perform useful tasks in many applications. Our main theoretical contribution is to prove that the training error of small networks converges linearly to a (nonnull) constant, of which we give a precise estimate. We verify this result on a neural network of 10 neurons simulating a Model Predictive Controller. We also observe that an upper bound of the generalization error follows a double-peak curve as the number of training data increases.

Keywords: Predictive control for nonlinear systems, Constrained control, Optimal control
Abstract: This paper investigates trajectory tracking of autonomous underwater vehicles (AUV) subject to thruster saturation and environmental disturbances. We propose a tube-based model predictive control (MPC) framework to robustly stabilize the AUV tracking error defined in the local frame. The essence of our design centers on leveraging a robust control contraction metric (RCCM) to construct a disturbance invariant set, ensuring bounded deviation between the actual and nominal system states under the RCCM-based feedback control law. Subsequently, an outer approximation of this RCCM-based invariant set is developed to design the tube cross-section and tighten the input constraint. The resulting RCCM-based tube MPC (RCCM-MPC) scheme is independent of the spatially varying metric, enhancing the computational efficiency of the proposed scheme. Under mild assumptions, we prove that the proposed RCCM-MPC scheme is recursively feasible and can asymptotically steer the AUV tracking error to a small region around the origin. Simulation results demonstrate the effectiveness of proposed RCCM-MPC approach.

Keywords: Predictive control for nonlinear systems, Machine learning, Stability of nonlinear systems
Abstract: This paper presents a novel learning-based distributed model predictive control strategy addressing the consensus problem in multi-agent systems with unknown state-dependent uncertainties. Assuming the availability of input-output data, a neural network-based model is used to describe state-dependent uncertainties. The Lipschitz constant of the neural network is estimated, and an upper bound on the prediction error of the neural network is derived based on this estimation. The multi-agent system is proven to be input-to-state stable concerning the prediction error. A numerical example is introduced to validate the theoretical results of the proposed controller. A comparison with other learning-based controllers illustrates the efficiency of the neural network-based controller. This research contributes to the advancement of learning-based model predictive control in scenarios involving multi-agent systems with inaccurate models.

Keywords: Constrained control, Lyapunov methods
Abstract: The increasing complexity of control systems necessitates control laws that guarantee safety w.r.t. complex combinations of constraints. In this letter, we propose a framework to describe compositional safety specifications with control barrier functions (CBFs). The specifications are formulated as Boolean compositions of state constraints, and we propose an algorithmic way to create a single continuously differentiable CBF that captures these constraints and enables safety-critical control. We describe the properties of the proposed CBF, and we demonstrate its efficacy by numerical simulations.

Keywords: Constrained control, Algebraic/geometric methods
Abstract: Given starting and ending positions and velocities, L₂ bounds on the acceleration and velocity, and the restriction to no more than two constant control inputs, this paper provides routines to compute the minimal-time path. Closed form solutions are provided for reaching a position in minimum time with and without a velocity bound, and for stopping at the goal position. A numeric solver is used to reach a goal position and velocity with no more than two constant control inputs. If a cruising phase at the terminal velocity is needed, this requires solving a non-linear equation with a single parameter. Code is provided on GitHub https://github.com/RoboticSwarmControl/MinTimeL2pathsConstraints/.

Keywords: Constrained control, Linear parameter-varying systems, Robust control
Abstract: We present a data-driven method to synthesize robust control invariant (RCI) sets for linear parameter-varying (LPV) systems subject to unknown but bounded disturbances. A finite-length data set consisting of state, input, and scheduling signal measurements is used to compute an RCI set and invariance-inducing controller, without identifying an LPV model of the system. We parameterize the RCI set as a configuration-constrained polytope whose facets have a fixed orientation and variable offset. This allows us to define the vertices of the polytopic set in terms of its offset. By exploiting this property, an RCI set and associated vertex control inputs are computed by solving a single linear programming (LP) problem, formulated based on a data-based invariance condition and system constraints. We illustrate the effectiveness of our approach via two numerical examples. The proposed method can generate RCI sets that are of comparable size to those obtained by a model-based method in which exact knowledge of the system matrices is assumed. We show that RCI sets can be synthesized even with a relatively small number of data samples, if the gathered data satisfy certain excitation conditions.

Keywords: Constrained control, Lyapunov methods, Optimization
Abstract: This paper introduces an approach for synthesizing feasible safety indices to derive safe control laws under state-dependent control spaces. The problem, referred to as Safety Index Synthesis (SIS), is challenging because it requires the existence of feasible control input in all states and leads to an infinite number of constraints. The proposed method leverages Positivstellensatz to formulate SIS as a nonlinear programming (NP) problem. We formally prove that the NP solutions yield safe control laws with two imperative guarantees: forward invariance within user-defined safe regions and finite-time convergence to those regions. A numerical study validates the effectiveness of our approach.

Keywords: Constrained control, Aerospace, Robust control
Abstract: This paper characterizes a class of constraint admissible positive invariant (CAPI) sets for vehicles with constrained nonlinear dynamics operating in SE(3). We present a robust linearization wherein the nonlinearities are eliminated using a combination of feedback-linearization and norm-bounded disturbances. We show that robust positive invariant sets for this robust linearization are positive invariant sets for the nonlinear dynamics. We present a trio of linear matrix inequalities for synthesizing the desired positive invariant sets. We present another pair of linear matrix inequalities which ensure that the feedback-linearizing controller is constraint admissible. We empirically show our candidate CAPI sets are invariant and constraint admissible through nonlinear simulation and compare with a candidate CAPI set synthesized using a non-robust linearization

Keywords: Constrained control, Optimal control, Autonomous robots
Abstract: Complex robotic systems require whole-body controllers to handle contact interactions, handle closed kinematic chains, and track task-space control objectives. However, for many applications, safety-critical controllers are essential to steer away from undesired robot configurations and prevent unsafe behaviors. A prime example is legged robotics, where we can have tasks such as balance control, regulation of torso orientation, and, most importantly, walking. As the coordination of multi-body systems is non-trivial, following a combination of those tasks might lead to configurations that are deemed dangerous, such as stepping on its support foot during walking, leaning the torso excessively, or producing excessive centroidal momentum, resulting in non-human-like walking. To address these challenges, we propose a formulation of an inverse dynamics control enhanced with control barrier functions that allow general higher-order relative degree safe sets for robotic systems with numerous degrees of freedom. Our approach utilizes a quadratic program that respects closed kinematic chains, minimizes the control objectives, and imposes desired constraints on the Zero Moment Point, friction cone, and torque. More importantly, it also ensures the forward invariance of a general user-defined high Relative-Degree safety set. We demonstrate the effectiveness of our method by applying it to the 3D biped robot Digit, both in simulation and with hardware experiments.

Keywords: Robotics, Identification for control, Modeling
Abstract: Soft robots are designed to be highly compliant to reduce the risk of injury or damage to humans and the environment. This compliance can lead to large deformations when handling payloads, significantly altering the operation of the robot. The stiffness of a soft robot, actively tuned or passively influenced by inputs (e.g., pneumatic pressures), is an important state variable, yet difficult to measure directly for the control of a soft robot. In this paper we propose a novel physics-informed approach to online estimation of the stiffness and shape of a soft manipulator under a payload, based on measurements that are readily available (e.g., positions of five points on the manipulator, or the position and orientation of the tip). The same approach is also adapted for estimating the payload when the stiffness is known. The proposed method is illustrated and supported with experimental results on a soft pneumatic actuator. In particular, it is shown to produce more accurate shape estimate than a commonly adopted piecewise constant curvature (PCC) model (which cannot produce a stiffness estimate), with an average error 57% smaller than the PCC method. The stiffness values estimated are also shown to be consistent and fall within the expected physical range.

Keywords: Feedback linearization, Nonlinear systems identification, Learning
Abstract: The emerging concept of morphological computation has shown that the complex dynamics of soft robots can be utilized as computational resources that facilitate their control. This work presents new insights into the capabilities of underactuated soft-bodied systems, as a class of partial feedback linearizable systems, for emulating dynamic responses of arbitrary nonlinear systems with stable limit cycles. An underactuated mass-spring-damper network (MSDN) was used as a physical reservoir. The dynamic equations are derived for the actuated and unactuated subsystems. A control law was derived to linearize the actuated subsystem, while the remaining portion is nonlinear and represents the internal dynamics. The stability of the linearized subsystem and corresponding zero dynamics were discussed for linear and nonlinear output functions. The effectiveness of the physical reservoir computing was demonstrated for emulating the Van der Pol (VDP) and Quadratic limit cycle (QLC) dynamics. A dynamics-informed learning process was established, which simulates the dynamics with a 1 kHz rate and trains the readout weights using the "teacher forcing" method subjected to the stability constraints with a 1 Hz rate. After training the readout weights, the feedback loop was closed, and the system was simulated where both oscillators were regenerated.

Keywords: Modeling, H-infinity control, Smart structures
Abstract: Dielectric actuators (DEAs) are lauded for their inherent softness, lightweight composition, and remarkable flexibility. These qualities empower DEAs with a unique capacity for finely tuned control, rendering them exceptionally suited for facilitating human-robot interactions. Their integration into prosthetic fingers introduces a unique advantage, enabling both delicate precision through soft actuation and broader manipulation via rigid activation. This paper presents a physics-based model for a prosthetic finger driven by DEAs, an effort enhanced by incorporation of feedback control mechanisms utilizing a self-sensing approach. This work not only underscores the potential of DEAs in creating lifelike human prostheses with enhanced dexterity, but also underscores their role in expanding the horizons of robotics for augmented human capabilities.

Keywords: Modeling, Identification, Mechatronics
Abstract: Twisted and coiled string (TCS) actuators are recently discovered motor-driven compliant actuators which generate actuation in two sequential phases, namely, twisting and coiling. TCS actuators' coiling-induced actuation is under-explored and exhibits unique characteristics: At a constant motor speed, twisting of strings generates contraction at increasing rate as the coil evolves from coil initiation stage to the coil formation stage. As the coiling process is discrete with individual coils forming sequentially, the aforementioned behavior generates non-smooth input-output correlation. To facilitate the application of TCS actuators in robotic and mechatronic devices, modeling and control of their behavior needs to be established. As a first step towards this objective, in this study, modeling and compensation for the unique non-smooth behavior of the TCS actuators in the coiling phase is presented. The proposed strategies leverage the geometric configuration of the coils. The model is formulated by relating the bias angle of each coil to the corresponding motor twist inputs. The proposed strategies were experimentally validated to perform satisfactorily: The model validation and control errors with the proposed strategies were 1.27%, and 1.11% of the overall actuation range, respectively.

Keywords: Mechatronics, Robotics, Machine learning
Abstract: This paper investigates the learning-based control of robotic systems driven by string-type artificial muscles. Due to the highly nonlinear dynamics of the actuators and the complicated mechanical structure, it is typically very challenging to design traditional model-based controllers that exhibit desired control performances. With the rapid development of machine learning techniques, deep reinforcement learning (DRL) algorithms have been utilized to control a variety of complex robotic systems. However, these DRL algorithms usually require a huge amount of training data and iterations to converge, which is generally unacceptable for the robotic systems of interest. Therefore, this paper designs an efficient learning-based control algorithm, aiming to improve the training efficiency for robotic systems driven by string-type artificial muscle actuators. Specially, two training improvements are proposed including imitation learning and data augmentation. This paper applies the proposed learning methods to three popular DRL algorithms and tests the control performances in three case studies using three string-type artificial muscle-driven robots, including a parallel robotic wrist, a two degrees-of-freedom (DOF) robotic eye and a robotic finger. Simulation results show that the proposed learning-based control methods significantly accelerate the convergence speed and improve the data efficiency in all the three case studies.

Keywords: Distributed parameter systems, Stability of nonlinear systems, Lyapunov methods
Abstract: The nonlinear Van der Pol oscillator is well-recognized for modeling limit cycles in electrical circuits. It was recently modified to a model whose parameters explicitly stood for the limit cycle amplitude and frequency. Due to this, the latter model has successfully been used in control engineering as a reference model for self-generation of limit cycles in the closed-loop. The popular Lotka-Volterra predator-prey partial differential equation (PDE) is another model which is capable of self-generating periodic orbits in infinite-dimensional setting. The present work aims to flavor the Van der Pol equation in PDE setting. It is shown that similar to the modified Van der Pol oscillator, its PDE-flavored model explicitly relies on the amplitude and frequency of the periodic orbit, which is self-generated by the model while also possessing a unique equilibrium in the origin similar to that of its finite-dimensional progenitor. Theoretical analysis is additionally supported by simulation results.

Keywords: Optimal control, Machine learning, Statistical learning
Abstract: This paper studies convergence rates for some value function approximations that arise in a collection of reproducing kernel Hilbert spaces (RKHS) {mathrm{ H}}(Omega). By casting an optimal control problem in a specific class of native spaces, strong rates of convergence are derived for the operator equation that enables offline approximations that appear in policy iteration. Explicit upper bounds on error in value function and control law approximations are derived in terms of power function mathcal {P}_{mathrm{ H,N}} for the space of finite dimensional approximants rm H_{N} in the native space {mathrm{ H}}(Omega). These bounds exhibit a distinctive geometric nature, refine and build upon some well-known, now classical results concerning the convergence of approximations of value functions.

Keywords: Cooperative control, Delay systems, Distributed parameter systems
Abstract: This paper presents a novel distributed encirclement control for first-order multi-agent systems that specifically considers the issue of communication delay. Unlike previous studies, this research analyzes its impact. One key aspect of our approach is the utilization of two distributed estimators affected by communication delay to accurately estimate the location of the geometric center of targets and to estimate the maximum distance of targets to the estimated geometric center of targets. The relative desired position of agents is achievable by employing the estimated maximum distance of targets to the estimated geometric center of targets. The employment of the estimated profile of targets and the relative desired position of agents allows for improved tracking and coordination among the agents, enhancing the effectiveness of the engaged encirclement control. To assess the stability of the closed-loop system, we employ the Lyapunov technique for analysis. Finally, we conduct simulations to evaluate the effectiveness of our proposed methodology.

Keywords: Distributed parameter systems, Delay systems, Control applications
Abstract: In this paper, we consider a 1D reaction-diffusion system with boundary input delay and propose a general method for studying the problem of prescribed-time output boundary stabilization. We first reformulate the system as a PDE-PDE cascade system (i.e., a cascade of a linear transport partial differential equation (PDE) with a linear reaction-diffusion PDE), where the transport equation represents the effect of the input delay. We then apply a time-varying infinite-dimensional backstepping transformation to convert the cascade system and the proposed observer system into two prescribed-time stable (PTS) target systems. The stability analysis is conducted on the target systems, and the desired stability property is transferred back to the closed-loop system and the error system using the inverse transformation. The effectiveness of the proposed approach is demonstrated through numerical simulations.

Keywords: Distributed parameter systems, Estimation, Modeling
Abstract: We propose a new approach for the estimation of the velocity field of normal flow partial differential equations by means of practical observers. In more detail, we suppose to know the evolution over time in each point of the domain of the solution of a normal flow equation subject to an unknown velocity field, which is assumed to be smooth but time-varying in general. Numerical results are presented to showcase the effectiveness of the proposed approach in one- and two-dimensional examples. In particular, the performance of the practical observer has been tested also to estimate the rate of spread of a fire propagation model based on the normal flow equation. Predicting wildfire evolution presents a multitude of challenges due to the variable nature of wildfires (influenced by a range of factors, from weather conditions and vegetation type to landscape topography), thus motivating the need to have at disposal an accurate estimate of the velocity vector to predict the evolution of the fire front.

Keywords: Agents-based systems, Networked control systems, Distributed control
Abstract: This paper proposes a framework for two- and three-dimensional distributed flocking control based on potential functions with ellipsoidal level sets. Contrary to most flocking frameworks in the existing literature, the presented approach allows for tuning the inter-agent distances in different directions individually. By defining a desired ellipsoidal lattice configuration based on an elliptic transformation matrix, the proposed flocking protocol can be formulated as a generalization of established spherical flocking protocols. Furthermore, it is shown how standard stability analysis can be extended seamlessly to the proposed framework without additional difficulties or assumptions. Simulation scenarios demonstrate the superior performance of the proposed scheme compared to existing ones.

Keywords: Kalman filtering, Linear parameter-varying systems, Energy systems
Abstract: We propose a short-term wind forecasting framework that enables model-based control systems to preemptively adapt ahead of atmospheric variations in improving turbine efficiency and reducing structural loads and failures. Our approach relies on a combination of linear stochastic estimation and Kalman filtering algorithms to assimilate and process real-time nacelle-mounted anemometer and surface air-pressure readings with the predictions of a stochastic reduced-order model of the hub-height velocity field. Our results serve as a proof of concept for a wind forecasting strategy based on ground-level pressure sensor measurements.

Keywords: Adaptive control, Energy systems
Abstract: Wind turbines degrade over time, resulting in varying structural, aeroelastic, and aerodynamic properties. In contrast, the turbine controller calibrations generally remain constant, leading to suboptimal performance and potential stability issues. The calibration of wind turbine controller parameters is therefore of high interest. To this end, several adaptive control schemes based on extremum seeking control (ESC) have been proposed in the literature. These schemes have been successfully employed to maximize turbine power capture by optimization of the Kw^2-type torque controller. In practice, ESC is performed using electrical generator power, which is easily obtained. This paper analyses the feasibility of torque gain optimization using aerodynamic and generator powers. It is shown that, unlike aerodynamic power, using the generator power objective limits the dither frequency to lower values, reducing the convergence rate unless the phase of the demodulation ESC signal is properly adjusted. A frequency-domain analysis of both systems shows distinct phase behavior, impacting ESC performance. A solution is proposed by constructing an objective measure based on an estimate of the aerodynamic power.

Keywords: Maritime control, Power systems, Optimal control
Abstract: Future Naval Microgrids (MGs) will include hybrid energy storage systems (ESS), including battery and supercapacitors to respond to emerging constant power loads (CPLs) and fluctuating pulsed power loads (PPLs). Voltage regulation of naval microgrids and power sharing among these resources become critical for success of a mission. This paper presents a novel control strategy using nonlinear model predictive controller embedded with a complex droop control architecture for voltage restoration and power sharing in medium voltage DC (MVDC) Naval MGs. The complex droop control ensures allocating supercapacitors (SCs) for high-frequency loads (i.e., PPLs), while battery energy storage system (BESS) and auxiliary generators share the steady-state load (i.e., CPL). Compared to state-of-the-art control of the naval ship MGs that relies on linear models, the proposed method incorporates the nonlinear behavior of the MGs in the closed-loop control framework via nonlinear model predictive control (NMPC). A reduced order representation of the MVDC dynamic is employed as the prediction model, augmented with a multi-objective, constraints-based, optimal control formulation. The results demonstrate the effectiveness of the proposed control framework for voltage restoration and power sharing of resources in naval MGs.

Keywords: Adaptive control, Fault accomodation, Electrical machine control
Abstract: This paper proposes a reinforcement learning-based method to maximize power generation for a direct-drive marine hydrokinetic turbine. A high levelized cost of energy (LCOE) is preventative in the widespread adoption of many marine energy conversion technologies. A straightforward way to reduce LCOE is to increase conversion efficiency and ensure maximum energy generation. The proposed method utilizes a damping control methodology, varying applied generator torque via a linear relationship between the applied damping coefficient and rotor speed. A state-action-reward-state-action (SARSA) algorithm has been used to learn the optimal con- trol action for a given flow velocity. The proposed SARSA methodology uses Gaussian radial basis functions to create a three-dimensional surface to estimate the relationship between damping coefficient, incoming flow velocity, and coefficient of power (Cp). With an example flow velocity profile derived from discharge data from a USGS test site while considering the effects of biofouling on the hydrodynamics of the turbine, the SARSA algorithm is compared against a baseline optimal tip speed ratio controller, where it has been found that the proposed RL generates 0.92% more energy than the baseline.

Keywords: Identification for control, H-infinity control
Abstract: The dynamic induction control wake mixing strategy has the potential to increase the energy yield of floating wind farms. These floating turbines will be subjected to surface waves, caused by the wind, and swell. When dynamic induction control is applied in open-loop, the effect of second-order wave forces and dynamic induction control on the thrust force can be out-of-phase and have destructive interference. In this work, we propose a method to synchronize the dynamic induction control input to the effect of the second-order wave forces. This is achieved by formulating the synchronization problem within an H-Infinity optimization framework and designing a controller that minimizes the difference between the effect of wave-induced thrust variation and thrust variation. Time domain simulations show that synchronization at a desired frequency can be achieved and that the overall performance of the dynamic induction control method can be enhanced.

Keywords: Control applications, Iterative learning control, Optimization
Abstract: This paper presents a novel data-driven wind farm control strategy to steer the wake. The approach uses the turbine power output and standard turbine measurement equipment as input, such as the nacelle anemometer and the wind vane. By designing the control strategy based on the measured data, sub-optimal yaw angle estimates from an analytical model can be compensated and inaccuracies induced by the turbine sensors can be corrected. Thereby advancing wake steering controllers to be independent of external sensors, such as a met mast or lidars. Measurements acquired from a wake steering field campaign are used to train the data-driven model and to evaluate the predictions of the model. The proposed data-driven approach highlights a consistent increase in the power gain compared to an analytical approach with a potential improvement ranging from 3.4% up to 8.0%.

Keywords: Cooperative control, Control of networks, Distributed control
Abstract: We investigate the consensus problem of multiagent systems with linear dynamics via a fully distributed event-triggered approach. To solve this problem, we present a fully distributed control strategy with an innovative event-triggered mechanism, where a clock variable is introduced to guarantee the existence of strictly positive minimum inter-event times (MIETs). Compared with existing results, the proposed event-triggered mechanism provides more flexibility for the clock variable, improving inter-event times. A positive lower bound of MIETs is derived, and is shown to be larger than those in previous works. The main results are illustrated with a numerical example.

Keywords: Cooperative control, Distributed control, Control of networks
Abstract: This paper studies the consensus problem of linear multi-agent systems (MASs) with relative output measurement on undirected graphs. Output-feedback protocols are designed for both leaderless and leader-follower consensus. Different from [1], dynamic event-triggered sampling strategies are integrated to reduce the communication burdens between agents. Compared with most of the existing results which are based absolute output, the proposed protocols are based on sampled relative output. Finally, a simulation example is provided to illustrate the theoretical results.

Keywords: Cooperative control, Game theory, Optimization
Abstract: This paper proposes an online cooperative pursuit strategy for multiple pursuers within an unbounded domain. One of the popular cooperative pursuit strategies is Voronoi-based pursuit, which is designed by minimizing the area of the evader's Voronoi cell. However, this pursuit strategy only works for bounded domains as the Voronoi cell is well-defined in a bounded environment. Instead, our strategy is built upon the minimization of the farthest squared distance between the evader and its safe-reachable set (SRS), which formulates the cooperative pursuit as an online concave optimization problem. To overcome the computational burden arising from solving the concave optimization problem, we linearize the SRS as a convex polygon by sampling vertices on the boundary of SRS, which transforms the concave optimization problem into a more tractable one. Moreover, we show that if the vertex number is large enough, our approach can guarantee a capture of the evader in finite time for cases with non-zero capture radii. Simulation results demonstrate the efficacy and efficiency of our approach.

Keywords: Cooperative control, Machine learning, Distributed control
Abstract: Effective multi-agent collaboration is imperative for solving complex, distributed problems. In this context, two key challenges must be addressed: first, autonomously identifying optimal objectives for collective outcomes; second, aligning these objectives among agents. Traditional frameworks, often reliant on centralized learning, struggle with scalability and efficiency in large multi-agent systems. To overcome these issues, we introduce a decentralized state-based value learning algorithm that enables agents to independently discover optimal states. Furthermore, we introduce a novel mechanism for multi-agent interaction, wherein less proficient agents follow and adopt policies from more experienced ones, thereby indirectly guiding their learning process. Our theoretical analysis shows that our approach leads decentralized agents to an optimal collective policy. Empirical experiments further demonstrate that our method outperforms existing decentralized state-based and action-based value learning strategies by effectively identifying and aligning optimal objectives.

Keywords: Cooperative control, Robotics, Optimization
Abstract: With the recent influx in demand for multi-robot systems throughout industry and academia, there is an increasing need for faster, robust, and generalizable path planning algorithms. Similarly, given the inherent connection between control algorithms and multi-robot path planners, there is in turn an increased demand for fast, efficient, and robust controllers. We propose a scalable joint path planning and control algorithm for multi-robot systems with constrained behaviours based on factor graph optimization. We demonstrate our algorithm on a series of hardware and simulated experiments. Our algorithm is consistently able to recover from disturbances and avoid obstacles while outperforming state-of-the-art methods in optimization time, path deviation, and inter-robot errors. See the code and supplementary video for experiments.

Keywords: Cooperative control, Robust adaptive control, Networked control systems
Abstract: Achieving resource-efficient robust tracking is critical for micro aerial robotics subject to external disturbances. In this paper, a novel distributed dual-layer adaptive event-triggered controller is proposed for cooperative formation tracking under unknown disturbances. Two different triggering sequences are designed to determine the appropriate instants for communication and control updates, thus greatly reducing the update frequency of the controller and communication consumption. Moreover, adaptive event-triggered conditions with heterogeneous parameters are well-designed to cope with unknown disturbances, while improving the flexibility of parameter selection and excluding Zeno behavior. By employing Lyapunov theory, new sufficient conditions for ensuring the feasibility of the proposed algorithm are rigorously derived to eliminate the conservatism. The performance of the proposed algorithm is illustrated through numerical validations of cooperative formation tracking under cluttered nonlinear perturbations, and a detailed comparison with existing algorithms is made to fully demonstrate the robustness and advantage of the proposed algorithm.

Keywords: Stability of linear systems, Linear systems, Linear parameter-varying systems
Abstract: In this paper (with its IP protected content), we introduce the concept of Convex Stability for analyzing the real state variable convergence issue (stability) of any Linear Time Invariant State Space (LTISS) system, x˙ = Ax, x(0) = xo. This is different from the current literature’s Hurwitz Stability concept because the definition of A^0 is different in the conditions we present for the Convex Stability of the matrix A. We propose Indices labeled, Transformation Allergic Indices, (TAIs) as real state variable convergence measures to guarantee the Convex Stability of the given real matrix A. For simpler exposition of this concept, in this paper, we present conditions for the Convex Stability property of a special class of X shaped real, square matrices, making them belonging to matrices with zero interconnections or interconnections-free (in the Qualitative ecological principles language). For such matrices, the proposed conditions of this paper do not require any eigenvalue concept at all. For the special case of the X matrices considered, the proposed conditions of Convex stability point out the misleading conclusions eigenvalue based conditions draw about the real state variable convergence of the same X matrix, thereby concluding that in general, Hurwitz stability conditions of any real matrix A (not just the X type matrices) can never guarantee Convex stability of the same real matrix A.

Keywords: Stability of linear systems, Linear systems, Linear parameter-varying systems
Abstract: In this paper, with its IP protected content, we introduce a new approach or philosophy (different from the Hurwitz Stability concept of the current literature) labeled as the Transformation Allergic (TA) Approach for analyzing the real state variable convergence of any real Linear Time Invariant State Space (LTISS) system, x˙ = Ax, x(0) = xo. The proposed TA Approach in turn defines a new concept labeled as Convex Stability of any real nth order real matrix which then hinges on the Convex Stability of a simple 2nd order matrix. The Convex Stability of a 2nd order real matrix is then analyzed by using recently introduced scalars labeled as Transformation Allergic Indices (TAIs). With the help of these TAIs, we identify a hidden instability labeled as Transformation Allergic Index Singularity which happens prematurely prior to the Determinantal Singularity predicted by Hurwitz Stable matrices. Then it is also shown that all Hurwitz Stable matrices turn out to be always Non-Convex. Thus it is emphasized that Non-Convexity itself is to be treated as Instability. Clearly, such premature Instability (Non-Convexity) needs to be considered as highly undesirable in safety critical LTISS systems such as those of autopilots for flight vehicles. Simple tests to identify Non-Convexity in any Hurwitz stable matrix are also presented.

Keywords: Stability of linear systems, LMIs, Large-scale systems
Abstract: In this paper we are concerned with stability of interconnected linear time invariant systems. It is shown that stability conditions involving specifically structured Lyapunov function candidates can be decomposed to a set of local and coupled stability-related conditions, without introducing any additional conservatism. Specific feature of the considered Lyapunov functions is that they include (higher order) time derivatives of the state vector. Decomposition of stability conditions is based on the notion of differential interconnection neutral functions, which are conceptually related to interconnection neutral supply functions from the theory of dissipative dynamical systems.

Keywords: Stability of linear systems, Switched systems, Uncertain systems
Abstract: This paper presents the first stabilization result for almost periodic piecewise linear systems (APPLSs) with both norm-bounded additive modeling uncertainty and dwell-time uncertainty. The technique employs a mixed-mode time-varying Lyapunov function to generate a sequence of controller gains that stabilize the uncertain APPLS. This modelling structure aligns with the R2R dry transfer of patterned two-dimensional (2D) materials, an emerging technology for continuous, chemical-free flexible material and device transfer. Thus, the proposed controller synthesis method provides stabilization guarantees for a novel modeling framework with applications in the expanding realm of advanced 2D device manufacturing.

Keywords: Stability of nonlinear systems, Lyapunov methods, Large-scale systems
Abstract: This paper proposes a module-based framework for analyzing and designing dynamical systems evolving on positive orthants and having stationary points in their interior. Dissipativity theory is a widely-used concept that allows one to assess a property of a system by aggregating dissipativity of its components. Popular storage functions and supply rates do apply to positive systems, but do not give information about the positivity. This paper employs non-vector space storage functions encompassing logarithmic functions to formulate input-to-state stability (ISS) with positivity guarantee, and shows how its dissipativity characterization can effectively address system interconnections. The developed supply rates and criteria are not only asymmetric to fit non-vector state spaces, but also carry equilibrium information which is inherent in positive systems exhibiting interior equilibria. Mathematical and practical examples illustrate that the proposed framework offers flexibility in applying to various system interconnections.

Keywords: Stability of hybrid systems, Autonomous robots, Autonomous systems
Abstract: We propose a hybrid feedback control strategy that safely steers a point-mass robot to a target location optimally from all initial conditions in the n-dimensional Euclidean space with a single spherical obstacle. The robot moves straight to the target when it has a clear line-of-sight to the target location. Otherwise, it engages in an optimal obstacle avoidance maneuver via the shortest path inside the cone enclosing the obstacle and having the robot's position as a vertex. The switching strategy that avoids the undesired equilibria, leading to global asymptotic stability (GAS) of the target location, relies on using two appropriately designed virtual destinations, ensuring control continuity and shortest path generation. Simulation results illustrating the effectiveness of the proposed approach are presented.

Keywords: Robust control, Delay systems, Networked control systems
Abstract: Our earlier work established a contention-resolving model predictive control (or MPC) framework to co-design priorities and control for networked control systems (or NCSs), assuming the system models are perfect. However, due to the differences between a model and the real world, there are inevitable perturbations. In this paper, we focus on perturbations of the predicted timings, which are rarely addressed in the literature, and present a robust MPC design to achieve guaranteed performance under such timing perturbations. We propose a state-feedback correction term with dynamic gain, added to the nominal contention-resolving MPC policy, which can eliminate the state deviation between perturbed and nominal systems at the predicted task completion time. Additionally, we identified the largest tolerable timing perturbation for such a robust MPC design. Under the tolerable timing perturbations, we analytically proved that the state deviation can be bounded by a forward invariant set (or FIS) for all time. The robust MPC policy can be then designed based on the FIS, such that perturbed system trajectories are guaranteed to satisfy all the original state and control constraints. The effectiveness of our proposed method is verified through simulation.

Keywords: Robust control, Linear systems, Optimization
Abstract: This paper focuses on the stabilization and regulation of linear systems affected by quantization in state-transition data and actuated input. The observed data are composed of tuples of current state, input, and the next state's interval ranges based on sensor quantization. Using an established characterization of input-logarithmically-quantized stabilization based on robustness to sector-bounded uncertainty, we formulate a nonconservative infinite-dimensional linear program that enforces superstabilization of all possible consistent systems under assumed priors. We solve this problem by posing a pair of exponentially-scaling linear programs, and demonstrate the success of our method on example quantized systems.

Keywords: Robust control, Optimal control, Optimization
Abstract: Frequency Response Function (FRF)-based control synthesis methods for Linear Time-Invariant (LTI) systems have been widely used in control theory and industry. Recently, there has been renewed interest in these methods, employing numerical optimization tools to enhance their performance. In this letter, we analyze the convergence properties of two data-driven frequencybased H2 and H∞ synthesis methods (Karimi and Kammer, 2017), (Schuchert et al., 2024) as the controller order increases. We demonstrate that, in the limit, an optimal controller can be designed using only the FRF, providing an estimate of the best achievable performance or enabling the design of an optimal high-order controller.

Keywords: Power systems, Robust control, Stability of nonlinear systems
Abstract: This paper considers a power transmission network model characterized by interconnected nonlinear swing dynamics on generator buses. At the steady state, frequencies across different buses synchronize to a common nominal value such as 50Hz or 60Hz, and power flows on transmission lines are at steady-state values. We assume that fast measurements of generator rotor angles are available. Our approach to frequency and angle control centers on equipping generator buses with large-scale batteries that are controllable on a fast timescale. We link angle based feedback linearization control with negative-imaginary systems theory. Angle based feedback controllers are designed for these large-scale batteries and can be implemented in a distributed manner with local information. Our analysis demonstrates the internal stability of the interconnection between the power transmission network and the angle based feedback controllers. This internal stability underscores the benefits of achieving frequency synchronization and preserving steady-state power flows within the network through the use of feedback controllers. This will enable transmission lines to be operated at maximum power capacity since stability robustness is ensured by the use of feedback controllers rather than conservative criteria such as the equal area criterion. Simulation results illustrate our results.

Keywords: Robust control, Stochastic optimal control, Optimization
Abstract: We consider a continuous-time continuous-space stochastic optimal control problem, where the controller lacks exact knowledge of the underlying diffusion process, relying instead on a finite set of historical disturbance trajectories. In situations where data collection is limited, the controller synthesized from empirical data may exhibit poor performance. To address this issue, we introduce a novel approach named Distributionally Robust Path Integral (DRPI). The proposed method employs distributionally robust optimization (DRO) to robustify the resulting policy against the unknown diffusion process. Notably, the DRPI scheme shows similarities with risk-sensitive control, which enables us to utilize the path integral control (PIC) framework as an efficient solution scheme. We derive theoretical performance guarantees for the DRPI scheme, which closely aligns with selecting a risk parameter in risk-sensitive control. We validate the efficacy of our scheme and showcase its superiority when compared to risk-neutral and risk-averse PIC policies in the absence of the true diffusion process.

Keywords: Robust control, Uncertain systems, Formal verification/synthesis
Abstract: Tracking controllers are often integrated into control systems to ensure their robustness against uncertainties and disturbances during trajectory following maneuvers, where design methods in the literature either lack formal guarantees or suffer from conservatism. In this paper, we propose a new tracking control approach for discrete-time nonlinear uncertain systems using set-based computations. In particular, we compute zonotopic backward reachable sets along prescribed nominal trajectories, and utilize such sets to synthesize tracking controllers that ensure safety and reachability in the presence of non-convex input/state constraints and uncertainties. We illustrate our approach through two numerical examples (Dubin's car and planar quadrotor).

Keywords: Kalman filtering, Estimation, Communication networks
Abstract: This paper proposes a data-driven attack strategy capable of circumventing sliding-window chi-square detectors in remote state estimation. The developed strategy is designed to operate based on only the intercepted output data from the plants, estimators, and anomaly detectors, without the knowledge of system parameters. Moreover, in scenarios where sufficient data has been collected before the occurrence of attacks, the proposed strategy exhibits optimality among all feasible attack policies using the same historical information. Through simulations, the effectiveness of the data-driven strategy is verified, with the stealthiness sustained by consistent empirical alarm rates.

Keywords: Kalman filtering, Estimation, Fault detection
Abstract: In chemical batch reactors, proper mixing is important for consistent process and product quality. Heel masses (mass that is left behind from batch to batch) can be a feature of process design, or may occur inadvertently. The current practice at the Hanford nuclear waste processing site is to include heel masses (up to 30% by volume) to reduce batch to batch variation of chemical waste that will be vitrified. To monitor this process, the proposed process uses laboratory measurements with a mass balance to propagate composition estimates from tank to tank. Incorporating in-line monitoring tools can improve the ability to detect faults when they occur in processing, particularly faults concerning tank transfer and heel masses. A simulation study is performed based on reported data on expected tank concentrations, sensor accuracies, and analytical laboratory accuracies. Observability of model states and parameters are shown. By incorporating in-line sensors, the mean absolute prediction error for the heel mass of kyanite, an insoluble solid, is improved by a factor of 3.88 over mass-balance model approaches in a simulation study including a fault in heel mass.

Keywords: Kalman filtering, Machine learning, Estimation
Abstract: We consider state estimation for a dynamical system that has unknown state-dependent dynamic process and measurement noise covariance matrices. When the noise covariances are state-dependent the typical Kalman filter (KF) fails to accurately estimate the states of the system. To estimate the states of such a system, we model the covariance matrices via the Wishart process and propose a novel variational Bayesian adaptive Kalman filter (VB-AKF). The proposed VB-AKF combines the variational Bayesian inference of the Wishart process with the KF. The resulting VB-AKF can estimate the states of the system together with the state-dependent dynamic process and measurement noise covariances. Through simulations, we show that the developed VB-AKF is effective and achieves satisfactory performance.

Keywords: Kalman filtering, Observers for nonlinear systems, Computational methods
Abstract: The weighted Kalman filter (WKF) can be perceived as a variant of the extended Kalman filter (EKF), which incorporates the so-called weighted linearisation. The price behind such an extension is expressed by an increased computational complexity, which aims at improving the overall convergence of the filter. Owing to the increased complexity, the WKF is in general not suitable for high-order nonlinear systems. To overcome this difficulty, the objective of this paper is to show that a closed-form expression of the WKF can be obtained for a specific class of nonlinear systems, i.e., those characterised by a quadratic state equation. This closed-form of the WKF scales well with the order of the system, enabling the application of the WKF to high-order nonlinear quadratic systems, as exemplified in the final part of the paper using a comprehensive Monte Carlo analysis applied to a large number of random systems.

Keywords: Kalman filtering, Sensor networks, Autonomous robots
Abstract: We consider two nonlinear state estimation problems in a setting where an extended Kalman filter receives measurements from two sets of sensors via two channels (2C). In the stochastic-2C problem, the channels drop measurements stochastically, whereas in 2C scheduling, the estimator chooses when to read each channel. In the first problem, we generalize linear-case 2C analysis to obtain -- for a given pair of channel arrival rates -- boundedness conditions for the trace of the error covariance, as well as a worst-case upper bound. For scheduling, an optimization problem is solved to find arrival rates that balance low channel usage with low trace bounds, and channels are read deterministically with the expected periods corresponding to these arrival rates. We validate both solutions in simulations for linear and nonlinear dynamics; as well as in a real experiment with an underwater robot whose position is being intermittently found in a UAV camera image.

Keywords: Optimal control, Estimation, Filtering
Abstract: We formulate the discrete-time inverse optimal control problem of inferring unknown parameters in the objective function of an optimal control problem from measurements of optimal states and controls as a nonlinear filtering problem. This formulation enables us to propose a novel extended Kalman filter (EKF) for solving inverse optimal control problems in a computationally efficient recursive online manner that requires only a single pass through the measurement data. Importantly, we show that the Jacobians required to implement our EKF can be computed efficiently by exploiting recent Pontryagin differentiable programming results, and that our consideration of an EKF enables the development of first-of-their-kind theoretical error guarantees for online inverse optimal control with noisy incomplete measurements. Our proposed EKF is shown to be significantly faster than an alternative unscented Kalman filter-based approach.

Keywords: Linear systems, LMIs, H-infinity control
Abstract: In this paper we propose a non-conservative static state-feedback controller synthesis method for discrete-time linear time-invariant systems. The synthesis goal is to render a closed-loop system dissipative with respect to a given generic unstructured quadratic supply function, while the system dynamics is partially represented by noisy state-input data. While the same problem has already been considered in the literature, the main novelty is the solution approach which is based on the dualization lemma to obtain solvable linear matrix inequalities for controller synthesis.

Keywords: Linear systems, Identification, Subspace methods
Abstract: In this paper, we revisit the classical problem of estimating the frequency response of an LTI system from noisy, non-periodic input-output data. Existing solutions fall into two categories: indirect methods, which compute the frequency response using identified models, and direct methods, which estimate the response directly from data. Direct methods bypass system identification, but have challenges when applied to noisy, finite-length datasets with unknown initial conditions. This paper proposes a new direct method that addresses these challenges, and provides an explicit formula for computing the frequency response. To develop this method, the paper leverages ideas from behavioral system theory and poses the problem as an optimization problem, whose objective is to minimize the projection of the solution onto the nullspace of an input-output data matrix. The paper also offers an alternative derivation of the formula, based on identifying an ARX model, thereby bridging the gap with classical indirect approaches. The proposed method is applied to experimental data collected from a DC motor, where we show that the proposed method outperforms other direct approaches based on Fourier transforms and low-rank approximations, and performs equally as good as direct subspace identification methods, even though no model class is prescribed.

Keywords: Linear systems
Abstract: The characterisation of single-input-single-output externally positive linear systems is considered. A complete characterisation of the class of externally positive second and a class of underdamped third order systems is given and connections to negative-imaginary systems are highlighted. It is shown that negative-imaginary systems have non-negative step responses, leading to a condition for external positivity based on negative imaginary systems theory. Finally, a counter-example is introduced for a recently developed numerical test for external positivity based upon linear matrix inequalities. These results extend the class of system for which external positivity can be verified, facilitating large-scale control and less conservative absolute stability analysis.

Keywords: Linear systems, Modeling, Adaptive control
Abstract: In the history of control theory, many methods have been proposed to simplify the procedure for determining the control inputs for complex dynamical systems. The most recent one of that series is “Model Free Control” which is very popular at the present time. The first three sections of the paper deal with these topics to provide the appropriate background to evaluate the main contribution of the paper contained in Section IV. This includes simulation studies on the following three dynamical systems: i) The roll stabilization of an aircraft, ii) The speed control of a motor, and iii) The stabilization of the longitudinal dynamics of an aircraft.

Both model-free control and model-based control were attempted in the above studies. While stability of the overall system and satisfaction of design specifications could be assured using model-based control, neither one was possible with model-free control. The results raise many questions which need to be addressed by those advocating model-free control. In particular, the limitations of the approach need to be clearly stated.