Hierarchical Model-Based Imitation Learning for Planning in Autonomous Driving
This work addresses autonomous driving planning in complex urban environments, offering a method that generalizes to unseen scenarios, though it is incremental as it builds on existing MGAIL techniques.
The paper tackles dense urban self-driving by applying hierarchical model-based generative adversarial imitation learning (MGAIL) to expert trajectories from over 100,000 miles of real driving, resulting in a steerable policy that generalizes to novel goals and approaches expert performance in closed-loop simulations.
We demonstrate the first large-scale application of model-based generative adversarial imitation learning (MGAIL) to the task of dense urban self-driving. We augment standard MGAIL using a hierarchical model to enable generalization to arbitrary goal routes, and measure performance using a closed-loop evaluation framework with simulated interactive agents. We train policies from expert trajectories collected from real vehicles driving over 100,000 miles in San Francisco, and demonstrate a steerable policy that can navigate robustly even in a zero-shot setting, generalizing to synthetic scenarios with novel goals that never occurred in real-world driving. We also demonstrate the importance of mixing closed-loop MGAIL losses with open-loop behavior cloning losses, and show our best policy approaches the performance of the expert. We evaluate our imitative model in both average and challenging scenarios, and show how it can serve as a useful prior to plan successful trajectories.