Q-learning with Long-term Action-space Shaping to Model Complex Behavior for Autonomous Lane Changes.

Gabriel Kalweit, Maria Hgle, Moritz Werling, Joschka Boedecker

Browse the full IROS paper archive.