Towards Lifelong Learning for Autonomous Driving

Lifelong learning approach is key feature which allows autonomous robots and vehicles to adapt its knowledge and behavior to changing circumstances over time, and ultimately to operate reliably and safely in our daily life. In this talk, we first present a dataset for autonomous driving, entirely based on ROS, recorded with up to eleven heterogeneous sensors including various cameras and lidars, a radar, an IMU (Inertial Measurement Unit), and a GPS/RTK (Global Positioning System / Real-Time Kinematic), in Montbeliard in France. This dataset captured many new research challenges for urban driving, such as sloping road, shared zone, diversion, etc., and as it includes daily and seasonal changes, it is especially suitable for long-term autonomy research. We then introduce two research projects (with CTU and Renault respectively), both of which are about how to use lifelong learning approach to deal with the challenges of autonomous driving in adverse weather conditions. Finally, we briefly introduce our participation in the Toyota Partner Robot joint research project.

August 8, 2019 (Thursday)
Room 205, Building E, Karlovo nám. 13


Dr. Zhi Yan

Dr. Zhi Yan is currently an Assistant Professor in the Distributed Artificial Intelligence and Knowledge Laboratory (CIAD) at the University of Technology of Belfort-Montbéliard (UTBM). From 2016 to 2017, he was a Postdoctoral Research Fellow in the Lincoln Centre for Autonomous Systems (L-CAS) at the University of Lincoln, working on the Horizon 2020 project FLOBOT. From 2013 to 2015, he was a Postdoctoral Research Fellow in the CAR Team at the IMT Lille Douai. In 2012, he received his Ph.D. from the Paris 8 University. His research interests are in autonomous driving, mobile robotics, and chronorobotics.