Project R-15208

AI-Driven Collective and Advanced Perception for Enhanced Decision-Making in Cooperative, Connected, and Automated Bus Mobility (Research)

The AICAP-Bus project aims to facilitate the seamless integration of autonomous buses into existing traffic systems. This goal necessitates enhancements in both the performance of autonomous vehicles (AVs) and their perception by road users. Positioned at the intersection of multiple disciplines, including human sciences and artificial intelligence, this project seeks to address these challenges holistically. By combining multi-sensor data with insights from hybrid intelligence and behavior models, the project's primary technical objective is to establish an advanced predictive control strategy for autonomous buses. This strategy will enable the vehicles to outperform current AVs in high-risk scenarios by anticipating environmental changes and emulating the driving style of seasoned drivers, thereby enhancing acceptance among various road users. The project seeks to advance autonomous vehicles in three key areas, ensuring their effective deployment in complex environments. A) Hardware improvements will enhance performance by improving environmental perception and reducing processing times, while maintaining financial feasibility through the use of cost-effective sensors. B) Software enhancements, leveraging cutting-edge AI models, hybrid intelligence, and behavior modeling, will surpass the capabilities of current AVs in interpreting user intentions and predicting environmental changes, thereby more effectively preventing risky situations. C) Finally, these advancements will contribute to a more positive user perception of autonomous technology, facilitated by the introduction of an advanced human-machine interface (HMI) both inside and outside the vehicle, along with transparent decision-making systems. The project will also prioritize the study of passenger perceptions, safety feelings, and the ethical dimensions of decision-making and continuous data acquisition.

10 September 2024 - 28 February 2025