Robust Control for Lane Keeping System Using a Linear Parameter Varying Approach with Scheduling Variables Reduction

This paper presents a robust controller using Linear Parameter Varying (LPV) model of lane-keeping system with parameter reduction. Both varying vehicle speed and roll motion on curved road influence the lateral model’s parameters, such as tire cornering stiffness. Thus, we use LPV technique to take variations into account in dynamics. However, multiple parameters lead high number scheduling variables cause massive computational complexity. In this paper, reduce complexity, Principal Component Analysis (PCA)-based reduction is performed obtain reduced tighter convex set. We designed feedback solving set Matrix Inequality (LMI). The effectiveness proposed validated full dynamics from CarSim an interchange road. From simulation, confirmed that method largely reduces offset error, compared other controllers based Time-Invariant (LTI) system.