[Paper] Data-driven Control for Human Driving Preferences

We have been developing a method that leverages both AI feature learning and model-based parametric MPC to learn driver model and preferences from driving data.

This work is a really nice effort from our research engineer, Flavia Sofia Acerbo, in collaboration between our R&D ADAS team at Siemens with experts from control systems (Prof. Jan Swevers, Mech Dept., KU Leuven) and computer vision (Prof. Tinne Tuytelaars, ESAT, KU Leuven). The works involve extensive testing and validation using driving simulator, (real-time) digital twin model building, and human-in-the-loop testing.

If you are interested, please check our paper here: https://lnkd.in/e5T-Wx_z