DeepRacer: Educational Autonomous Racing Platform for Experimentation with Sim2Real Reinforcement Learning

TL;DR

DeepRacer is a scalable platform enabling autonomous racing using reinforcement learning with raw camera inputs, demonstrating effective sim2real transfer, robust policy training, and comprehensive evaluation methods.

Contribution

The paper introduces a novel autonomous racing platform that leverages deep reinforcement learning with raw camera data, showcasing successful sim2real transfer and large-scale deployment.

Findings

First successful large-scale deployment of deep RL on a robotic control agent using raw images.

Effective sim2real transfer achieved without additional tuning in the physical environment.

Development of a robust evaluation method to determine training completion.

Abstract

DeepRacer is a platform for end-to-end experimentation with RL and can be used to systematically investigate the key challenges in developing intelligent control systems. Using the platform, we demonstrate how a 1/18th scale car can learn to drive autonomously using RL with a monocular camera. It is trained in simulation with no additional tuning in physical world and demonstrates: 1) formulation and solution of a robust reinforcement learning algorithm, 2) narrowing the reality gap through joint perception and dynamics, 3) distributed on-demand compute architecture for training optimal policies, and 4) a robust evaluation method to identify when to stop training. It is the first successful large-scale deployment of deep reinforcement learning on a robotic control agent that uses only raw camera images as observations and a model-free learning method to perform robust path planning. We open source our code and video demo on GitHub: https://git.io/fjxoJ.