Keywords: Energy Systems, Fault detection/accomodation, Learning
Abstract: Batteries are integral parts of many modern applications, and over the last decade, battery technologies have transformed grid storage, portable electronics, and transportation. As these technologies penetrate more critical applications, their prognostics and diagnostics become more important. First, battery production is not clean, and using them for longer is crucial to their promise as an environmentally friendly solution [1]. Second, these batteries can pose a serious hazard if faults go undetected, leading to thermal runaway and fire [2, 11].

This workshop is based on the previous work in the following areas: optimal EIS data collection [5-6], and processing EIS data by modeling it or deconvoluting the data into a distribution of relaxation times [2,7], online fault detection [3,12], feature selection for diagnostics [MS, 8, 10, 12], as well as learning battery models [4,9,10]. Time-domain data to extract features for machine learning [11]. Methods such as Federated learning, which have emerged as a promising paradigm for data-driven modeling, have applications in battery health monitoring for automotive and medium-duty applications, where sharing raw data may be impractical due to privacy, bandwidth, and ownership constraints [14-15].

This workshop presents new theoretical and experimental approaches for prognostics and diagnostics of batteries. The workshop materials range from introductory material on the possibilities of using machine learning for more efficient data collection to data preprocessing and, eventually, to their use for prognostics and health management (PHM) in AI settings. In this workshop, we will address several challenges and opportunities with AI tools for PHM: (i) provide an overview of the type of data and data collection, (ii) present techniques to model the batteries for feature selection, (iii) use AI for learning battery models from data, and (iv) use AI tools for PHM. This workshop features presenters from the academy and industry. Also, they are from different geographical locations, including Europe, Canada, and the USA.

Keywords: Autonomous systems, Robotics, Robust control
Abstract: Autonomous systems are rapidly gaining popularity across safety-critical domains such as transportation (both aerial and ground), logistics, environmental monitoring, and space exploration. However, ensuring their safe, robust, and efficient operation remains a significant challenge, especially in dynamic and uncertain environments with limited onboard computational resources.

This workshop will explore the latest advancements in control strategies for autonomous systems, with an emphasis on the integration of safety guarantees, computational efficiency, and robustness to uncertainty. Particular attention will be given to methods that enable real-time onboard implementation, as well as the integration of low-level control with high-level planning and decision-making. Key topics include, but are not limited to: • Safety-critical and constraint-aware control for autonomous systems • Robust and adaptive control with safety and performance guarantees • Learning-based control for autonomous systems • Computationally efficient planning and control for real-time onboard implementation • Integration of planning, control, and high-level decision-making • Applications to autonomous aerospace, ground, surface, and underwater vehicles, and beyond

This workshop aims to bridge the gap between theoretical developments and real-world deployment by bringing together researchers and practitioners from control theory, machine learning, robotics, and autonomous systems. Through invited talks and interactive discussions, the workshop will foster deeper insights into safe, robust, and efficient autonomy, identify key open challenges, and promote new research directions in this rapidly evolving field. Profile

Keywords: Predictive control, Optimization, Software tools
Abstract: Model Predictive Control is a well-established technique for controlling (possibly) nonlinear systems that has predictive ability and can cope with constraints. It has a strong track record in chemical engineering. More recent advances in problem formulations and high-performance solvers have expanded MPC's applicability to domains such as power systems, mechatronics, and robotics, with complex dynamics and high sampling rates.

In this tutorial, participants will engage in hands-on exploration of MPC applied to mechatronic systems using Impact. Impact is a flexible toolchain for the specification, prototyping, and deployment of optimal control problems (OCP) and model predictive control (MPC) strategies, with automatic generation of deployable artifacts. Impact is built on top of the Rockit toolkit for OCPs and the highly popular CasADi symbolic framework for numerical optimization.

The key contribution of the toolchain, and this workshop, is to reduce the engineering complexity of MPC implementations by providing: 1. an intuitive symbolic tool with abstraction of technical details that are cumbersome to implement, and 2. flexible integration with state-of-the-art numerical optimization solvers, such as acados, Fatrop, and GRAMPC, for rapid prototyping and high-performance embedded deployment. Impact is written in Python and offers bindings for MATLAB. The generated artifacts can be executed from C/C++, Python, MATLAB, Simulink, and ROS 2 environments, and easily deployed in simulation and on hardware. Workshop exercises will be in Python. Attendees will gain practical experience and can adopt the presented open-source software frameworks in their research and applications.

Keywords: Reinforcement learning, Learning
Abstract: This half-day workshop, entitled Advances in Reinforcement Learning for Control Applications, covers timely technical issues at the intersection of reinforcement learning, control theory, safety, and emerging real-world applications. The proposed program runs from 13:30 to 17:00 and is organized around five invited talks that together highlight both fundamental advances and application-driven challenges in modern learning-based control.

A central theme of the workshop is reliability and safety in reinforcement learning. As reinforcement learning is increasingly considered for deployment in safety-critical domains, such as autonomous systems, medical treatment design, manufacturing, and cyber-physical systems, it is no longer sufficient to focus only on reward maximization. Instead, there is a growing need for methods that explicitly incorporate safety constraints, uncertainty handling, interpretability, and domain knowledge. This workshop addresses these issues from complementary theoretical and practical perspectives.

Keywords: Energy Storage, Energy Systems, Power systems
Abstract: The world is on the cusp of a new era of electrification across different sectors of industry and economy. A key technology driving this transformation is lithium-ion batteries. As the best available power source, lithium-ion batteries provide high energy/power density and long cycle life. As they find every-growing use in electric vehicles, electric aircraft, grid storage and autonomous platforms, demands for their performance and safety have been rising. Systems and control theory can play a key role in meeting the needs to advance the application of battery systems, resulting in provable progresses.

This tutorial-style workshop is designed to provide a deep, structured introduction to the state of the art and new frontiers in battery modeling, monitoring and control, with a focus on integrating physical insights and control-theoretic rigor with practical implementation. The workshop will be particularly relevant to researchers and practitioners within the systems and control community looking to expand their research towards developing and applying control-theoretic methods for lithium-ion batteries and emerging battery-powered systems.