1 paper · 1 filter
Kanghoon Lee, Jiachen Li, David Isele +3
Although deep reinforcement learning (DRL) has shown promising results for autonomous navigation in interactive traffic scenarios, existing work typically adopts a fixed behavior p…