Benchmarking Empirical and Learning-Based Approaches for Feedforward Steering Control in Autonomous Racing

Summary

A new research paper systematically benchmarked two learning-based and two empirical feedforward steering controllers for autonomous racing, introducing a new Empirical Hysteresis Dynamics (EHD) formulation. The study utilized a high-fidelity simulation framework based on the real-world Abu Dhabi Autonomous Racing League competition. While learning-based controllers showed lower prediction errors in open-loop evaluation, the proposed EHD approach achieved the best overall closed-loop robustness and lap times. Why it matters: This research highlights the critical importance of evaluating control strategies within a complete software stack for autonomous racing, directly informing the development for competitions like the AADRL.