Drift control for cornering maneuver of autonomous vehicles

Abstract

Expert drivers have the ability to perform high side-slip angle maneuvers, like drifting, during racing to minimize lap time or avoid obstacles. Designing planning and control algorithms for autonomous drift maneuvers, however, is challenging because of the high lateral motion and nearly full saturation of rear tires. In this paper, the authors propose a complete path planning and motion control framework to plan and track a reference drift trajectory along a sharp bend in a track. The path planner divides the path horizon into three regions, finds a path using different planning sub-modules, and then concatenates the solutions to generate the reference trajectory. The controller then applies a mixed open-loop and closed-loop scheme to track the reference trajectory. We validate the planning and control algorithms in simulation using a high-fidelity model in Simulink/Carsim, and through experimentation using 1/10 scale Radio-Control (RC) vehicle.