Adversarial Imitation Learning via Random Search in Lane Change Decision-Making
This work addresses decision-making for autonomous driving in multi-lane highways, but it appears incremental as it applies an existing optimization approach to a specific domain.
The paper tackled the problem of coordinating ADAS functions for autonomous lane change decisions by proposing a derivative-free optimization imitation learning method, achieving desired performance in simulation-based evaluations.
As the advanced driver assistance system (ADAS) functions become more sophisticated, the strategies that properly coordinate interaction and communication among the ADAS functions are required for autonomous driving. This paper proposes a derivative-free optimization based imitation learning method for the decision maker that coordinates the proper ADAS functions. The proposed method is able to make decisions in multi-lane highways timely with the LIDAR data. The simulation-based evaluation verifies that the proposed method presents desired performance.