Traffic-Aware Autonomous Driving with Differentiable Traffic Simulation

While there have been advancements in autonomous driving control and traffic simulation, there have been little to no works exploring the unification of both with deep learning. Works in both areas seem to focus on entirely different exclusive problems, yet traffic and driving have inherent semantic relations in the real world. In this paper, we present Traffic-Aware Autonomous Driving (TrAAD), a generalizable distillation-style method for traffic-informed imitation learning that directly optimizes a autonomous driving policy for the overall benefit of faster traffic flow and lower energy consumption. We capitalize on improving the arbitrarily defined supervision of speed control in imitation learning systems, as most driving research focus on perception and steering. Moreover, our method addresses the lack of co-simulation between traffic and driving simulators and lays groundwork for directly involving traffic simulation with autonomous driving in future work. Our results show that, with information from traffic simulation involved in supervision of imitation learning methods, an autonomous vehicle can learn how to accelerate in a fashion that is beneficial for traffic flow and overall energy consumption for all nearby vehicles.

Download Appendix

Average velocity of traffic flow over time for each of the 26 CARLA Leaderboard Test Scenarios. We show the traffic flow of LBC (baseline) versus our method for each of the 26 Test Scenarios over 1000 timesteps. For most scenarios, with the exception of scenarios 8, 16, and 22, our method improves the local traffic flow by increasing average velocity.

@INPROCEEDINGS{zheng2023trafficdriving,
  author={Zheng, Laura Y. and Son, Sanghyun and Lin, Ming C.},
  booktitle={2023 International Conference on Robotics and Automation (ICRA)}, 
  title={Traffic-Aware Autonomous Driving with Differentiable Traffic Simulation}, 
  year={2023}
}