Optimizing Renault’s embedded AI
Philippe THIERION, ADAS Development Engineer, Groupe Renault
“Above all, the Confiance.ai program indicated that internal AI developments were all at the level of research on the issue, while providing additional knowledge on specific points not previously considered. It also indicated that the flow of porting algorithms to target was by no means a solved subject and perfectly mastered in general by the community.”
Renault’s participation in the Confiance.ai program has led to significant advances in the development of its embedded AI systems. Optimizing hardware resources and improving ADAS and driver monitoring are at the heart of the issues addressed.
AI adoption challenges specific to Renault
The main challenge in the field of embedded AI is to make optimum and safe use of available HW resources, so as to rapidly embed a large number of algorithms in vehicles.
Products and services affected by the Confiance.ai program
The development of ADAS and driver monitoring will be positively impacted by the very definition of flow followed by Confiance.ai. Various software bricks, whose importance has been highlighted by Confiance.ai, are currently under development.
Confiance.ai’s contribution to the Renault Group’s strategy and challenges
The flow defined by the Confiance.ai program has enabled us to assess the key areas where development efforts should be focused, particularly in view of the rapid evolution of technology and the issue of embedded AI.
Competitive advantages developed with Confiance.ai
Controlling the AI development flow enabled Renault to assess feasibility in relation to the competition, and cost in relation to suppliers.
Challenges and conditions for success for the next phase of the Confiance.ai program
Target porting techniques are moving fast, for example to optimize the footprint of neural networks. These new techniques deserve to be studied.
Example of current or future implementation/deployment
The ongoing deployment of embedded AI systems using camera images has been greatly simplified by the lessons learned from the project.
Accelerating AI Act compliance with the Confiance.ai program
For Renault, Confiance.ai can provide knowledge and awareness of compliance.
