A Cyber-Physical System represents a system of computational elements which are collaborating in order to ensure the coordination, monitoring and control of physical processes. Such systems are typically designed as a network of interacting embedded computers with sensors as physical inputs and actuators as outputs. While an important effort is being made in the domain of Computer Sciences to enhance the design of embedded hardware, communication networks, real-time scheduling algorithms, etc., it is a challenging Control Theory problem to understand the interaction between discrete control algorithms and physical processes in a networked/embedded configuration. From the Control Theory point of view, the analysis and design of Cyber-Physical Systems implies the study of complex, hybrid models. We encounter dynamical systems involving switching in vector fields, as for example in the models describing scheduling protocols or representing their interaction with control algorithms and physical processes. We also encounter them in models describing the switching among several energy consumption (in the framework of multi-energy systems) or resource sharing modes, and more generally in systems involving jointly automatons and differential equations. Furthermore, Cyber-physical systems imply the study of dynamical systems with discontinuous state evolutions, representing the dynamics of sensor and control data in communication channels with impulsive changes at transmission times. Such classes of hybrid dynamical systems are very interesting from a theoretic point of view since the study of their basic properties (in terms of stability, stabilization, observability, etc.) is a largely open field. The aim of the Young Research project ROCC-SYS is to enhance the fundamental research towards the development of new analysis and control design methodologies for Cyber-physical systems via a Robust Control theory approach. New stability analysis methods and control design techniques will be developed for Hybrid Dynamical Systems. Tools for reliability study of Cyber-physical systems (in terms of Lyapunov stability) will be provided through the use of Hybrid System Theory. With respect to the current call of proposals, the ROCC-SYS consolidates the activities on hybrid dynamical systems and networked/embedded control of 4 young permanent researchers of the team SYNER of LAGIS. ROCC-SYS combines both theoretical research (analysis and control design methodologies for Cyber-physical systems based on hybrid system theory) and applied research, with the final goal being the development of software packages (Matlab programs) for the analysis and design of Cyber-physical systems and experimental validation.

Fuel Cell Systems (FCS) appear nowadays to be a promising and alternative energy source to face economic and environmental challenges of modern society. However, even if this technology is close to being competitive, it is not yet ready to be considered for large scale industrial deployment: FCS still must be optimized, particularly by increasing their limited lifespan. This involves not only a better understanding but also requires emulating the behavior of the whole system. Additionally, a new area of science and technology emerges: prognostic of FCS is a field of scientific and industrial developments that should be increased. This is the aim of the project: the partners feel a clear motivation for such a research and propose to develop intelligent Prognostics and Health Management (PHM) methods in order to assess the health state of Proton Exchange Membrane Fuel Cell systems (PEMFC), and predict its remaining useful life. To our knowledge, this is an original and still unexplored field: although the research in PHM is continuously increasing, one can note that efforts in the area do not deal with Fuel Cells applications. This may be (partially) explained by the lack of knowledge on the behavior of those systems. Indeed, it is difficult to develop prognostics tools that take into account the inherent uncertainty of not well understood failure mechanisms. In addition, prognostic presents deployment challenges: it is difficult to know how to set a prognostic tool, as well as is there is no systematic way to judge from it without waiting for the irreversible deterioration of the equipment. Following all this, scientific objectives of the project are defined as follows: - develop approaches for reliable prognostics of PEMFC stacks; - facilitate their implementation; in order to move towards a generic approach compatible with industrial constraints. Three related axis of developments are expected. 1. The first axis deals with handling the prognostic process. Here, the aim is not only to be able to estimate the remaining useful life of the fuel cell (provide prognostics estimates), but also to improve the capabilities of the developed prognostics tools by quantifying and controlling their inherent error of estimates. Several approaches will be considered: a model-based prognostic tool based on a Bond Graph approach with parametric uncertainty, and data-oriented prognostics tools by extending "signal" and "connexionist" approaches (like neural networks and neuro-fuzzy systems). The development of an hybrid approach will also be addressed. 2. The second axis scopes to enhance the applicability of prognostics tools. The purpose of this part of works is to look for solutions that enable systematizing the building of a prognostics system while reducing the influence of arbitrary human choices, as well as reducing its learning and/or parameterization times. Several options will be studied: the construction of adaptive and parsimonious systems, the definition of relevant "learning patterns", the identification of singular representatives learning samples, among others. 3. The third axis will focus on the industrial adoption process for fuel cells. The diffusion and transfer of results to industrials cannot be conceived without a precise knowledge of their expectations. It is thus necessary to understand the diffusion process of the Fuel Cell technology in the industrial world, and to identify the corresponding bottlenecks, other than those related to technological aspects. The project consortium has a competence in humanities and social sciences that will help to develop this axis.

Statistical methods have become more and more popular in signal and image processing over the past decades. These methods have been able to tackle various applications such as speech recognition, object tracking, image segmentation or restoration, classification, clustering, etc. We propose here to investigate the use of Bayesian nonparametric methods in statistical signal and image processing. Similarly to Bayesian parametric methods, this set of methods is concerned with the elicitation of prior and computation of posterior distributions, but now on infinite-dimensional parameter spaces. Although these methods have become very popular in statistics and machine learning over the last 15 years, their potential is however largely underexploited in signal and image processing. The aim of the overall project, which gathers researchers in applied probabilities, statistics, machine learning and signal and image processing, is to develop a new framework for the statistical signal and image processing communities. Based on results from statistics and machine learning we aim at defining new models, methods and algorithms for statistical signal and image processing. Applications to hyperspectral image analysis, image segmentation, GPS localization, image restoration or space-time tomographic reconstruction will allow various concrete illustrations of the theoretical advances and validation on real data coming from realistic contexts.

When a train crosses a border, it will need to change its onboard signalling system for example, which will generate an important financial cost. Interoperability of the rail system within Europe is therefore key to its competitiveness. It aims at creating a rail network allowing transport that is safe, compliant with the required performance level of the lines, and which does not necessitate train transfers. This requires the compliance with a set of rules, of technical and operational conditions which ensure that the essential requirements are met. The present project aims at contributing to the validation and implementation of a European system for railway signalling called ERTMS “European Rail Traffic Management System”. The management of railway signalling in ERTMS is based on local rules pertaining to each country and not on global rules. This makes it difficult to evaluate the system in terms of safety. Thus, one of the main objectives of this study is to supply methodological tools for the evaluation of the global consistency between the specification and the operating rules, with regard to safety. This issue is crucial and yet it has scarcely been covered by scientific literature. A formal representation of the ERTMS specifications will be provided, which will enable the validation various systems through automatically generated test scenarios. • These systems will be modelled and studied in order to establish, if possible without having to test, whether they are compliant with ERTMS • The national rules complementing ERTMS will be included in the study. Achieving interoperability through ERTMS requires that all parties have the same understanding of the technical specifications for interoperability. Using a model featuring rigorous semantics will help identifying and clarifying ambiguousness in the specifications. A first step will therefore be to build a formal model based on complex requirements taken out of some rules. It will then be possible to analyze in details a European specification in the face of national operating rules, for example with regard to the execution of a Movement authority (MA). An actual specification, proposed by the railway company, will be analyzed in work-package 1 (The ERA or EPSF will be asked to provide an initial specification). The French Public Railway Safety Authority (EPSF) is competent in railway security matters whiles being independent from railway operators. A second step will provide a study of methodological and software tools in the literature. Formal models will be used to determine whether a given scenario meets the specifications. The respect of the European system requirement specifications (SRS) and of national safety requirements will be studied, on the level of models as well as analysis tools. In a last step, a study will be based on tests on an ERTMS simulation tool compliant with the official specifications. The possibility of making the two work together in an integrated approach will be studied.