Residual Policy Learning for Vehicle Control of Autonomous Racing Cars
This work addresses the problem of improving practical applicability and performance for autonomous racing car controllers, representing an incremental advancement in the field.
The paper tackles the challenge of developing high-performance vehicle controllers for autonomous racing cars by proposing a residual policy learning approach that combines classical controllers with learned residual controllers. The result is an average lap time reduction of 4.55% compared to classical controllers on simulated racetracks, with gains also observed on unknown tracks.
The development of vehicle controllers for autonomous racing is challenging because racing cars operate at their physical driving limit. Prompted by the demand for improved performance, autonomous racing research has seen the proliferation of machine learning-based controllers. While these approaches show competitive performance, their practical applicability is often limited. Residual policy learning promises to mitigate this drawback by combining classical controllers with learned residual controllers. The critical advantage of residual controllers is their high adaptability parallel to the classical controller's stable behavior. We propose a residual vehicle controller for autonomous racing cars that learns to amend a classical controller for the path-following of racing lines. In an extensive study, performance gains of our approach are evaluated for a simulated car of the F1TENTH autonomous racing series. The evaluation for twelve replicated real-world racetracks shows that the residual controller reduces lap times by an average of 4.55 % compared to a classical controller and even enables lap time gains on unknown racetracks.