Roboracer

The 'Roboracer' dataset/benchmark consists of a framework for evaluating autonomous racing performance using 1:10-scaled racing vehicles, specifically designed to assess the effectiveness of reinforcement learning algorithms in real-world scenarios.

Papers using Roboracer (3)

Efficient Real-World Autonomous Racing via Attenuated Residual Policy Optimization2026Adaptive Control in Autonomous Driving via Real-Time Recurrent RL2026Context-Aware Model-Based Reinforcement Learning for Autonomous Racing2025