Autonomous Drifting of a 1:5 Scale Model Car

Abstract

Current automotive safety systems restrict the vehicle to the linear region of operation where the sideslip angle is small. Recent research in the field has discovered that drifting possesses unstable equilibria in which the vehicle is controllable even after the handling limits in the linear region have been exceeded.This thesis presents the design and simulation of a feedback linearization controller that, by using yaw rate as input to controlling the sideslip angle, is able to find the equilibrium point corresponding to the initial velocity and the desired yaw rate. Simulation results show that the controller is able to achieve a yaw rate within 5 degrees of the desired yaw rate. It is demonstrated that utilization of drifting techniques increases the maneuverability of the vehicle compared to normal cornering. Based on the successful handling of coupling in actuator authority at high angles of sideslip, feedback linearization as a control design tool is recommended for further development of controllers in the LocalHawk project.The LocalBug simulator has been improved by the addition of a dc motor model that includes selection of front, rear and four wheel drive. Measurement of the moment of inertia of LocalBug and recording of true noise data, which is added to the simulator output, has increased the fidelity of the simulator. Validation of the simulator shows that the simulation results largely is in agreement with logged test data, except for the case of hard breaking where the simulation model is inclined to experience a spin.