Robust Adaptive Triple-Step Control for Lane-Keeping of Autonomous Ground Vehicles

Active steering control for lane keeping is a significant technique applied for autonomous ground vehicles (AGVs). This paper addresses a problem of high-accuracy control for lane keeping of AGVs subject to time-varying vehicle speed and parameter uncertainties. A robust adaptive triple-step nonlinear control scheme is presented. In this strategy, a nonlinear reference generator is designed to generate a virtual reference signal. Then, a robust triple-step nonlinear controller with adaptive update laws is then proposed for lane keeping of AGVs. It is shown by the Lyapunov stability analysis that the lateral offset of AGV is input-to-state (ISS) in the presence of bounded perturbation of tire stiffness coefficient. The effectiveness of the approach is verified on a high-fidelity full-vehicle model constructed in veDYNA and MATLAB/Simulink.