ROBUST PATH FOLLOWING CONTROL FOR INDEPENDENTLY ACTUATED AUTONOMOUS GROUND VEHICLES WITH TRANSIENT PERFORMANCE IMPROVEMENT

Abstract

This thesis addresses the path following control of independently actuated (IA) autonomous ground vehicles (AGVs), and focuses on the constraint control and transient performance improvement. The control aim is to inhibit overshoots and eliminate steady-state errors in path following manoeuvres, and thus to reduce the possibilities of AGVs to surpass safe driving bounds. Centering on this focus, this thesis proposes a novel path-following modeling method, a novel output constraint control approach, and two novel nonlinear control strategies for transient performance improvement. The contributions of this study lie on the following five aspects: 1) A novel definition for the desired heading is proposed using the sideslip-angle compensation, to reduce the steady-state errors in path following caused by the non-zero sideslip-angle and the large/changing path curvature; 2) A novel output constraint control approach using hyperbolic projection technique is proposed to compactly restrain the lateral offset, and thus to prevent the vehicle from surpassing safe lanes; 3) A novel integral sliding mode (ISM)-based composite nonlinear feedback (CNF) technique is proposed to improve the transient performance of path following, considering the system uncertainties, disturbances and actuator saturations; 4) A novel disturbance observer-based CNF technique is proposed to improve the transient performance for path following with differential drive assisted steering (DDAS) when the active steering system is in complete failure. Corresponding comparative simulations based on CarSim-Simulink were conducted to verify the proposed modeling and control strategies in 1)-4); 5) Experiments have been conducted to verify the existence and effectiveness of the DDAS mechanism.