Home // SnT // News & E... // PhD Defence: Minimizing Supervision for Vision-Based Perception and Control in Autonomous Driving

PhD Defence: Minimizing Supervision for Vision-Based Perception and Control in Autonomous Driving

twitter linkedin facebook email this page
Add to calendar
Speaker: François Robinet (SEDAN group)
Event date: Tuesday, 04 October 2022 02:00 pm - 04:00 pm

You are cordially invited to attend the PhD Defence of Mr François ROBINET on 4th  October 2022 at 14:00 in room JFK-E00-004.

Members of the defense committee:

  • Prof. Dr Radu STATE, University of Luxembourg, Chairman
  • Prof. Dr Djamila AOUADA, University of Luxembourg, Deputy Chairman
  • Prof. Dr Raphaël FRANK, University of Luxembourg, Supervisor
  • Prof. Dr Christian MÜLLER, German Research Centre for Artificial Intelligence, Member
  • Dr Christian HUNDT, Nvidia Germany, Member
  • Dr Geoffrey NICHIL, Foyer Luxembourg, Expert in an Advisory Capacity
  • Michel ETIENNE, Foyer Luxembourg, Expert in an Advisory Capacity

Abstract:

The research presented in this dissertation focuses on reducing the need for supervision in two tasks related to autonomous driving: end-to-end steering and free space segmentation. 

For end-to-end steering, we devise a new regularization technique which relies on pixel-relevance heatmaps to force the steering model to focus on lane markings. This improves performance across a variety of offline metrics. In relation to this work, we publicly release the RoboBus dataset, which consists of extensive driving data recorded using a commercial bus on a cross-border public transport route on the Luxembourgish-French border. 

We also tackle pseudo-supervised free space segmentation from three different angles: (1) we propose a Stochastic Co-Teaching training scheme that explicitly attempts to filter out the noise in pseudo-labels, (2) we study the impact of self-training and of different data augmentation techniques, (3) we devise a novel pseudo-label generation method based on road plane distance estimation from approximate depth maps. 

Finally, we investigate semi-supervised free space estimation and find that combining our techniques with a restricted subset of labeled samples results in substantial improvements in IoU, Precision and Recall.