Sampling-based Bi-level Optimization aided by Behaviour Cloning for Autonomous Driving

Abstract

Autonomous driving has a natural bi-level structure. The upper behavioural layer aims to provide appropriate lane change, speeding up, and braking decisions to optimize a given driving task. The upper layer can only indirectly influence the driving efficiency through the lower-level trajectory planner, which takes in the behavioural inputs to produce motion commands for the controller. Existing sampling-based approaches do not fully exploit the strong coupling between the behavioural and planning layer. On the other hand, Reinforcement Learning (RL) can learn a behavioural layer while incorporating feedback from the lower-level planner. However, purely data-driven approaches often fail regarding safety metrics in dense and rash traffic environments. This thesis presents a novel alternative; a parameterized bi-level optimization that jointly computes the optimal behavioural decisions and the resulting downstream trajectory. The proposed approach runs in real-time using a custom Graphics Processing Unit (GPU)-accelerated batch optimizer and a Conditional Variational Autoencoder (CVAE) learnt warm-start strategy and extensive experiments on challenging traffic scenarios show that it outperforms state-of-the-art Model Predictive Control (MPC) and RL approaches regarding collision rate while being competitive in driving efficiency.