A real-world data-driven simulator for interaction-aware autonomous driving benchmarking
funded by DAAD Programmes for Project-Related Personal Exchange (PPP) with Scotland.
PI: Dr. Yu Feng, Dr. Jianglin Lan
As autonomous vehicles (AVs) increasingly operate in mixed traffic environments alongside human road users, ensuring safe and efficient interactions among all road users is becoming critically important. However, developing a socially acceptable interaction-aware autonomous driving (IAAD) system presents significant challenges. These challenges arise from the need for advanced methods to capture human-AV interactions, socially-compliant vehicle control algorithms, and realistic simulators for AV design and testing, all of which depend on extensive real-world data. Addressing these challenges requires a robust cross-disciplinary approach supported by long-term national and international collaborations. To this end, we are applying for the Project-Related Personal Exchange (PPP) program to facilitate a partnership between the Technical University of Munich (TUM) and the University of Glasgow (UofG). This collaboration aims to achieve three key objectives: building a comprehensive real-world dataset using roadside LiDAR, developing models to accurately capture human behaviour to enhance AV decision-making and control, and creating a data-driven simulator for evaluating and advancing socially-compliant autonomous driving technologies. Ultimately, this project will lay the groundwork for a mixed-reality platform that seamlessly integrates real and virtual traffic participants, promoting a truly safe, efficient, and socially-compliant intelligent transportation system.