Partial End-to-end Reinforcement Learning for Robustness Against Modelling Error in Autonomous Racing

TL;DR

This paper introduces a partial end-to-end reinforcement learning approach for autonomous racing that improves robustness to vehicle model mismatches by decoupling planning and control, combining RL with classical controllers.

Contribution

The paper proposes a novel partial end-to-end RL framework that separates planning from control, enhancing robustness against model errors in autonomous racing.

Findings

Outperforms standard end-to-end RL in robustness to model mismatches.

Combines RL-based trajectory planning with classical control for improved reliability.

Demonstrates effectiveness in autonomous racing scenarios.

Abstract

In this paper, we address the issue of increasing the performance of reinforcement learning (RL) solutions for autonomous racing cars when navigating under conditions where practical vehicle modelling errors (commonly known as \emph{model mismatches}) are present. To address this challenge, we propose a partial end-to-end algorithm that decouples the planning and control tasks. Within this framework, an RL agent generates a trajectory comprising a path and velocity, which is subsequently tracked using a pure pursuit steering controller and a proportional velocity controller, respectively. In contrast, many current learning-based (i.e., reinforcement and imitation learning) algorithms utilise an end-to-end approach whereby a deep neural network directly maps from sensor data to control commands. By leveraging the robustness of a classical controller, our partial end-to-end driving algorithm exhibits better robustness towards model mismatches than standard end-to-end algorithms.