Research activities

My research activities are about the safe design of embedded real-time systems. I am interested in both theoretical and practical problematics concerning the specification, the modeling and the validation of these systems. My works have been first conducted within the Ep-Atr and Espresso research groups at IRISA in Rennes (France), from 1999 to 2005. They rely on the multi-clock synchronous approach, also termed polychronous. Since the end of 2005, I have joined the DaRT research group at LIFL and INRIA Lille Nord Europe labs in Villeneuve d'Ascq (France), in which I keep on applying the synchronous approach to the safe design of high-performance embedded systems.

Parallel and distributed embedded systems

  • Model-driven design and validation of high-Performance SoCs: I am currently working on the reliable design of high-performance embedded systems, characterized by intensive data processing. Such systems are encountered in domains such as multimedia, radars, sonars, software radio, etc. Programmable and high-performance processor architectures are increasingly needed for the design of these systems. The non functional constraints, typically real-time properties, must finely take into account the specificity of these architectures for performance reasons. Yet, they condition the safety, which is often essential in the target applications. My work aims at defining a solution taking into account all the crucial constraints related to these systems and favoring their costless validation. For that, I am considering a combination of two models of computation (MoCs): the repetitive MoC and the synchronous reactive MoC. The former MoC enables to express the regularity of data parallelism in considered systems, and of the deployment of the systems on parallel multiprocessor architectures (with multiple clocks). The projection of this deployment in the synchronous reactive MoC allows one to get access to a wide range of formal proof tools, that facilitating the guarantee of performances and the validity of considered design choices.

  • Synchronous modeling of asynchrony in distributed systems: I am also interested in the design of distributed embedded systems using the synchronous multi-clock model of the Signal language. The main objective of this work is to show the capability of the synchronous approach to apply formal techniques and tools that guarantee the reliability of systems including asynchronous mechanisms. This work also aims to pragmatically help engineers to understand the manipulated formal concepts and apply them easily in order to solve system correctness issues encountered in practice.

Formal verification

  • Static Analysis of Signal Programs: I am taking a look at the static analysis of programs specified in the synchronous language Signal in order to improve the verification of numerical properties. My last result on this topic is the implementation of a library of functions that allow one to manipulate a generalized form of interval decision diagrams (IDD) initially proposed by Strehl et al. This library has been co-implemented with one of my students (G. Atigossou).

Safety-critical real-time and embedded systems

  • Design and Validation of Avionic Applications: In a very near past, I used to work intensively on the polychronous modeling of real-time systems, and in particular on avionic systems. I considered the Integrated Modular Avionics (IMA) architecture paradigm and its dedicated ARINC standard. I have defined a library of component models for this purpose. This library is available today in the design environment of the Signal language (Polychrony), for the description, the simulation and the validation of avionic applications. It has been experimented in the modeling and temporal analysis of a case study from Airbus France. It has also served in other works on the re-engineering of real-time Java programs and on model-driven engineering.

Projects

Below are the projects in the framework of which my research activities have been conducted:

  • The french ANR project "Famous" (Flot de modélisation et de conception rapide pour les systèmes dynamiquement reconfigurables) - starting in october 2009...
  • The french Collaborative Research Initiative Triade (Combining models of computation for the design of real-time and embedded applications) - still running...
  • The franco-tunisian project Ksour (Modèle, vérification et intégration de MPPSoC sur des architectures reconfigurables) - still running...
  • The french ANR project OpenEmbedd (An open-source platform for Model-Driven Engineering of Real-Time and Embedded systems) - now finished!
  • The franco-british project Modeasy (MOdel-Driven dEsign for Automotive Safety embedded sYstems) - now finished!
  • The European project IST SafeAir (Advanced Design Tools for Aircraft Systems and Airborne Software) - now finished!
  • The french RNTL project Expresso (Real-time Java for critical embedded systems) - now finished!


Tool development

Gaspard environment

I am participating to the development of the Gaspard design environment, which is dedicated to the model-driven design of high-performance embedded systems.

Polychrony/SME environment

I defined a library of polychronous component models for the design of avionic applications. The models are based on the avionic APEX-ARINC 653 standard. A presentation of this library is available here (contact: L. Besnard).