Simulation of Autonomous Vehicule Safety

The SVA Project, within the framework of the Autonomous Vehicle Plan of the New Industrial France, aims to meet the challenge posed by the complexity of the safety demonstration of the autonomous vehicle through the use of digital simulation. This complexity is related to both the high number of different situations that drivers encounter on the road, and their uncertainties, and the embedded technologies themselves. It indeed makes validation by real use testing extremely expensive if not impossible for some cases.


  • Provide builders and suppliers with a methodology, a platform and simulation tools for both the design and the validation of safe autonomous vehicles.
  • Specify, adapt or develop models of vehicle components and of its environment in order to simulate the behavior of the vehicle in the of occurrence of a failure of one of its components and/or to simulate the operational impact of an external disturbances (lack of clear surface road markings, rain, glare, etc.).


To interpret its environment, the autonomous vehicle will have sensors (radars, cameras, etc.), map information, data streaming from the infrastructure or other vehicles. Like any other system that can create the potential for risk prone events, it must be designed to ensure the safety of its occupants and that of other road users. For this, some components will need to be redundant. Furthermore, the vehicle could experience a degradation of the efficiency of its sensors, for example due to rain, a glare or smears. It could also have difficulties in interpreting the information supplied by the sensors, for example if the road surface markings have been erased. Making the autonomous vehicle safe in light of these external interferences without malfunction is a real challenge.

The project will address in particular:

  • The identification of the limits of current methodologies, including ISO 26262, as well as proposals for the development of the highlighted methods.The provision of methods and modeling tools that will be used in particular to construct a simulation platform of an autonomous vehicle in its environment.
  • The construction of test scenarios describing driving situations, that integrate external disturbances, in a format usable by the simulation platform.

Doctoral thesis supported by the project

  • Application des méthodes formelles à la maîtrise du développement des systèmes embarqués sur le véhicule autonome (ENS Cachan – Université Paris-Salay / Laboratoire Spécification et Vérification)
  • Formal models for the conceptualization and the characterization of use cases for the autonomous vehicule (Université de Versailles Saint-Quentine-en-Yvelines / Laboratoire DAVID / STIC)


The project results will be methods and software tools directly exploitable by our industrial partners in the automotive sector, with extensions planned for public transport and industrial vehicles.

Autonomous Transport
Industrial Partners
All4Tec ANSYS Apsys Assystem Continental Optis Oktal PSA Peugeot Citroën Renault Sector Valeo
Academic Partners
Project Manager
Jean van Frank