Enhancing the quality of Valeo’s perception systems for vehicle autonomy
Sonia KHATCHADOURIAN, R&I project manager, Valeo
“The Confiance.ai program is a research program based on industrial use cases. This enables us to get a direct grip on the reality of what is expected of trust in AI-based solutions over time, for a variety of fields. Within Valeo, we have been able to acquire and implement skills to guarantee a certain level of trust, particularly in the use of data. Finally, the standardization/AI Act aspect in the program lays the foundation for compliance with future AI regulations in the industry.”
By contributing to the Confiance.ai program, Valeo has improved the quality and reliability of its perception systems for autonomous vehicles by integrating trusted components and developing robust methodologies. These advances support the company’s competitiveness and ensure better compliance with future regulations.
AI adoption challenges specific to Valeo
Valeo needs to develop high-quality perception systems as the basis for vehicle empowerment. This includes system design, data management, model creation and deployment in vehicles.
Products and services affected by the Confiance.ai program
Confiance.ai has impacted the production of quality datasets, data management and dataset updating. The program has also improved perception solutions for autonomous vehicles and competitiveness through AI.
Confiance.ai’s contribution to Valeo’s strategy and challenges
Confiance.ai contributed to the evaluation of the training and test datasets and to compliance with the AI Act.
Competitive advantages developed with Confiance.ai
Valeo’s contribution to Confiance.ai has enabled Valeo’s teams to build skills on AI trust and its implications, anticipate the AI Act and improve data management.
Challenges and conditions for success for the next phase of the Confiance.ai program
For Valeo, Confiance.ai must continue to mature and become more permanent, so that its results are well established. Certain aspects, such as foundation models, remain to be dealt with, as the subject is so vast.
Example of current or future implementation/deployment
Valeo is currently rolling out part of the end-to-end methodology on a specific use case and assessing the results. The current scope is limited to data, but the company would like to broaden it to take advantage of Confiance.ai’s advances in both systems and on-board capabilities.
Accelerating AI Act compliance with the Confiance.ai program
Thanks to the program, Valeo can anticipate the application of the AI Act.