Last updated on February 28, 2015. This conference program is tentative and subject to change

The algorithm is based on modelling the known coding regions of a genome as a set of sample paths of a stochastic process, and the known non-coding regions of the same genome as a set of sample paths of another stochastic process. Then an ORF (Open Reading Frame) is classified as being either a coding region or a non-coding region based on likelihood estimation. Initially, each stochastic process is modelled as a fifth-order Markov process. Then a further reduction in the size of the state space is realized by observing that different strings can have `memories' of different lengths (which is the rationale for the name).

The 4M algorithm is applied to 70 or so genomes from both bacterial genomes and archaea, and its performance is compared to that of Glimmer-2, one of the most widely used algorithms. The 4M algorithm consistently matches or exceeds the performance of Glimmer-2 in the test cases. The size of the state space used by the 4M algorithm is a few hundred states, compared with 16,384 for Glimmer-2. Moreover, since the 4M algorithm is based on standard methods in stastical analysis, the significance of the various tests performed can be estimated precisely.

The method is based on two main ideas: * use of Pade approximation to transform the system into some singularly perturbed finite-dimensional system, for which robust dichotomy has to be checked; * recursive applications of Generalized Kalman-Yakubovich-Popov (KYP) lemma to characterize by an LMI the previous property.

In detail, we first introduce the reachability definition and a necessary condition for reachability. We then fully explore the "pattern reachability" problem, which is the problem of reaching, for every possible zero pattern (i.e. a specific choice of which entries are positive and which entries are zero), at least one nonnegative state vector endowed with this zero pattern.

Finally, for the class of n-dimensional single-input systems, switching among n possible subsystems, necessary and sufficient conditions for reachability are given.

Our new scheme maintains a fixed number of candidate paths in a history, each identified by an optimal subset of estimated mode probabilities. The memory requirements of our filter are fixed in time and can be varied by the user to achieve the desired accuracy.

Computer simulations are given to demonstrate performance of the Gaussian-mixture algorithm described, against the IMM.

Recently, two signal-design-algorithms for active failure detection were introduced: one that computes optimal continuous detection signals for continuous systems (CS) and the other that designs optimal piecewise-constant signal inputs for sampled-data systems (SDS). In some applications simple piecewise constant inputs are sought. In this paper, we combine the key ideas of the two previous algorithms and present an algorithm that finds optimal piecewise-constant signals for continuous systems. This modified algorithm provides a suboptimal detection signal inputs for continuous systems and would appear to be greatly faster than the original CS algorithm. We also compare the three algorithms and the corresponding optimal signals through a computational experiment.

Keywords: Fault diagnosis problem, nonlinear systems, reduced-order observer, algebraic observer.

We study the linear combinations of the process and its shifts that produce a process independent of the input. The set of all such linear combinations, called the orthogonalizers, has a module structure and under identifiability conditions completely specifies the deterministic part of the ARMAX system. Computing a module basis for the orthogonalizers is a deterministic identification problem.

We propose an ARMAX identification algorithm, which has three steps: first compute the deterministic part of the system via the orthogonalizers, then the AR part, which also has a module structure, and finally the MA part.

We show that the notion of (epsilon,delta)- approximate (bi)simulation can be thought of as a generalization or relaxation of the earlier work on delta- approximate (bi)simulation by Girard and Pappas. We demonstrate the link between reachability verification and approximate (bi)simulation, and we also provide a characterization of (bi)simulation relations using a tool similar to the (bi)simulation function.

Approximate synchronization can be thought of as a generalization of synchronization of transition systems in the usual sense. In fact, the usual synchronization and interleaving synchronization are two special cases of the notion of approximate synchronization developed in this paper. Furthermore, we present a result on the compositional properties of the approximate (bi)simulation with respect to the approximate synchronization.

Using the same technique, if the system is ISS, under some conditions on the system equations and on the smooth ISS Lyapunov function, then the state x of the system converges to zero for any L^{p} input.

In this paper, we propose a novel adaptive controller, referred to as safe adaptive controller, which is composed of a supervisor and two candidate controllers. One candidate controller is a LTI model reference controller that can always guarantee system stability, while the other is a model reference adaptive controller that can tune its parameters to counteract the changes in the plant. The supervisor evaluates the performance of the two candidate controllers without using any plant model information and activates the candidate controller leading to better transient response. It is shown that the system stability is guaranteed as long as the non-adaptive candidate controller is stabilizing. Simulation results are presented to demonstrate the proposed safe adaptive controller is able to achieve better performance than the two candidate controllers.

Our method builds upon the explicitly parametrized control formulae that we introduced in our earlier work on non-adaptive backstepping control for PDEs. These formulae allow us to develop tunable controllers that avoid solving Riccati or Bezout equations at each time step.

We demonstrate the method using an aircraft fault detection scenario and show that the new method significantly reduces the bound on the probability of error when compared to a manually generated identification sequence and a fuel-optimal sequence.

We consider a multiterminal system designed for efficiently estimating a random parameter according to the MMSE criterion. The analysis is limited to scalar quantizers followed by a joint entropy encoder, and it is performed in the high-resolution regime where the problem can be easier mathematically tackled.

Current implementations of the proposed approach become faster than existing methods for large problems. Extensions of this method are proposed that make the computational requirements lower than those of existing approaches in all cases, while allowing for efficient parallelisation and bounded memory usage.

This paper concentrates on the simple and intuitive integral quadratic distance measure. For the special case of a Dirac mixture with equally weighted components, closed-form solutions for special types of densities like uniform and Gaussian densities are obtained. Closed-form solution of the given optimization problem is not possible in general. Hence, another key contribution is an efficient solution procedure for arbitrary true densities based on a homotopy continuation approach.

In contrast to standard Monte Carlo techniques like particle filters that are based on random sampling, the proposed approach is deterministic and ensures an optimal approximation with respect to a given distance measure. In addition, the number of required components (particles) can easily be deduced by application of the proposed distance measure. The resulting approximations can be used as basis for recursive nonlinear filtering mechanism alternative to Monte Carlo methods.

It is shown that the lateral speed dynamics and the yaw rate dynamics can be decoupled by feeding back longitudinal speed, yaw rate and lateral acceleration measurements: lateral speed measurements are not required. The yaw rate tracking error dynamics follow a second order reference model with arbitrary poles, while the lateral speed dynamics tend exponentially to zero with a vehicle-dependent time constant and lateral acceleration tends to be proportional to the yaw rate.

Simulations on a nonlinear third order single track model show significant improvements in the closed loop behaviour: larger stability regions, larger bandwidth, resonances suppression, improved manoeuvrability. A key feature of the input-output decoupling control is the improved comfort since both the lateral speed and the phase lag between lateral acceleration and yaw rate are greatly reduced.

As illustrative example, the control system design of a reverse osmosis desalination plant is used. Simulation results are satisfactory and show that in many cases, as for example this desalination plant, multi-loop control with several controllers, which have been obtained by join multi-objective optimization, perform as good as more complex controllers but with less implementation effort.

In our previous work, we examined the benefits of vehicle control on the message delay. Specifically, we obtained optimal delay scaling where each message was required to be picked up and delivered by the same vehicle. Motivated by application to wireless networks, in this paper we remove this restriction and allow the vehicles to relay message between them. Specifically, we consider two relay methods. The first requires vehicles to relay messages directly to each other using a synchronous rendezvous schedule while the other utilizes an infinite capacity depot to store relayed messages. Under both relay models, we characterize the minimal delay scaling which demonstrates that relaying helps in reducing delay further. Surprisingly, the optimal delay scaling is achieved with only one relay per message.

We note that our results naturally apply to the classical vehicle routing setup as well as to a wireless communication network. Specifically, our results suggest that the delay reduction can be very significant in a controlled relay network.

Exponential asymptotic stability indicates the ability of the algorithm to track modest movements of the source.

Efficient simulation-based methods using convolution particle filters are proposed. The regularization properties of these filters is well suited, given the context of parameter estimation. Firstly the usual non Bayesian statistical estimates are considered: the conditional least squares estimate (CLSE) and the maximum likelihood estimate (MLE). Secondly, in a Bayesian context, a Monte Carlo type method is presented. Finally we present a simulated case study.

We examine two models of the tokamak-and-plasma system, one assuming the plasma has mass, the other assuming zero mass. Although the plasma with mass model is more correct, the massless model is most often used in control analyses. We find that answers to questions regarding vertical stability, with or without feedback, depend on whether the plasma is assumed to have mass or not. We provide examples where analyses conducted using a massless plasma model can reach erroneous conclusions