Real-Time and Embedded Systems team

Control systems with critical functions or Quality of Services (QoS) constraints are supported by computers. A correct behaviour of the system does not only depend on correctness of results but also on time instants at which they are produced. Such a dual requirement characterizes real-time systems. A computerized control system must be reactive towards the controlled systems or QoS metrics to guarantee. In most of implementations, the control software is decomposed in a set of parallel tasks upon one or several processors. Before using the real-time system, designers must verify that temporal constraints will be respected throughout the system life. To ensure that task deadlines will be met, scheduling algorithms must be designed and configured according to the system specifications and the environment of the controlled processes. Finally, the system is validated by a detailed scheduling analysis in order to prove the temporal constraints will be met.

Scientific problems under study

  • On-line / Off-line scheduling,
  • Design of Real-Time Systems,
  • Performance analysis and Quality of Service,
  • Real-Time Application Modeling,
  • Analysis of Schedule Properties.

Application Domains

  • Embedded Systems (avionic, automotive...),
  • Industrial Process Control,
  • Multiprocessor systems,
  • Heterogeneous Distributed Systems.

Set up methods, techniques and approaches

  • Algorithm Analysis, Computational Complexity, Approximation, Combinatorial Algorithms,
  • Modeling and Validation using finite automata, Petri Nets, Markovian processes, discrete geometry, Quality of Service Analysis using metrics,
  • Combinatorial optimization: mathematical programming, Branch and Bound, metaheuristics (genetic algorithms, simulated annealing, ...),
  • Simulation.