Benjamin CATHELINEAU

In cyber-physical systems, the presence of various electronic components and software controlling and coordinating the operation of elements interacting with the physical world (sensors, actuators, etc.) induces a wide variety of possible faults and disturbances. This context makes the analysis of various types of system-level properties, including fault tolerance and dependability, extremely complex.
In general, these problems affect a wide variety of application domains, such as aeronautics, automotive, intelligent management of electrical distribution networks, industrial control and automatic regulation, etc.
The objective here is to develop methods and tools allowing the analysis of the impact at the system level of faults, disturbances, failures, this analysis being feasible from the first stages of the design flow thanks to simulable models allowing reasoning on the requirements at a very high level of abstraction. This involves offering solutions for automatically testing the impact of various fault and disturbance models in nominal execution scenarios, and providing diagnostic assistance if necessary in the form of recommendations for modifications to architectural choices. This being done with a great flexibility as for the interchangeability in simulation of the models of subsystems with respect to the expressiveness of the faults considered.