General description

PRISSMA, second pillar of the Grand Challenge on AI-based systems certification

The PRISSMA project is the response proposed by the autonomous mobility sector to Pillar 2 issued by the Grand Défi in partnership with the Ministry of Ecological Transition and Solidarity (in particular through the DGITM and the DGEC), on the securing, the reliability, and eventually the certification of AI-based systems. 

This project aims at proposing a platform that will allow to lift the technological barriers preventing the deployment of secure AI-based systems, and to integrate all the elements necessary for the realization of the homologation activities of the autonomous vehicle and its validation in its environment for a given use case.

We can list three main objectives for this project.

  1. Identify the safety and security objectives for AI-based autonomous mobility systems and develop comprehensive reliability validation processes for commercial operation of autonomous mobility services.
  2. Ensure the availability of shared concepts to address the complexity of AI-based autonomous mobility systems that can be used internationally.
  3. Participate in the implementation of prerequisites allowing France to position itself at the European level to host one of the Testing Facilities for autonomous mobility that will be developed in the coming years.

The PRISSMA project focuses on two main use cases, namely shuttles and automated robots (Droids).

An ambitious project

  • PRISSMA deals with AI in the approval of Automated Mobility by considering the whole chain: Vehicle(s), Infrastructure (augmented or not) and Supervision.
  • PRISSMA will propose a methodological framework complementary to the existing certification framework related to AI issues, with the objective of setting up a reference system or a label.
  • Based either on an existing homologation strategy (shuttles) or by generating a new homologation strategy (droids), PRISSMA aims to :
    • extract a maximum of information from the tests in simulated environment, in particular for the identification of the critical scenarios,
    • enrich tests in controlled and real environments, especially in relation to supervision and infrastructures,
    • adapt and develop test and simulation technologies/tools/means for the implementation of proofs of concept (POC) validating the adequacy of the process and tools in relation to the needs of the approval strategy.