A Super-human Vision-based Reinforcement Learning Agent for Autonomous Racing in Gran Turismo
This addresses the challenge of developing vision-based autonomous racing agents without external instrumentation, which is important for real-world robotics applications, though it builds incrementally on prior work that used global features.
The paper tackles the problem of creating a super-human autonomous racing agent using only local sensor inputs, such as ego-centric camera pixels and on-board car data, and achieves performance that outperforms the best human drivers in time trial races on multiple tracks and cars in Gran Turismo 7.
Racing autonomous cars faster than the best human drivers has been a longstanding grand challenge for the fields of Artificial Intelligence and robotics. Recently, an end-to-end deep reinforcement learning agent met this challenge in a high-fidelity racing simulator, Gran Turismo. However, this agent relied on global features that require instrumentation external to the car. This paper introduces, to the best of our knowledge, the first super-human car racing agent whose sensor input is purely local to the car, namely pixels from an ego-centric camera view and quantities that can be sensed from on-board the car, such as the car's velocity. By leveraging global features only at training time, the learned agent is able to outperform the best human drivers in time trial (one car on the track at a time) races using only local input features. The resulting agent is evaluated in Gran Turismo 7 on multiple tracks and cars. Detailed ablation experiments demonstrate the agent's strong reliance on visual inputs, making it the first vision-based super-human car racing agent.