Varying Zonotopic tube RMPC with switching logic for lateral path tracking of autonomous vehicle. Issue 7 (May 2022)

• Record Type: Journal Article
• Authors: Zheng, Hao; Zheng, Ling; Li, Yinong; Wang, Kan; Zhang, Ziwei; Ding, Minghui
• Is Part Of: Journal of the Franklin Institute. Volume 359:Issue 7(2022)
• Page Start: 2759
• Page End: 2787
• Publication Date: 2022-05
• Subjects: Science -- Periodicals; Technology -- Periodicals; Patents -- United States -- Periodicals; 505
• DOI: 10.1016/j.jfranklin.2022.03.011
• Languages: English
• ISSNs: 0016-0032

Abstract

Highlights: A complete framework for practical application of AV systems is proposed based on a novel scheme of Varying Zonotopic TRMPC with Switching Logic (SVTMPC). The new update mechanism of the present nominal state can determine the unknown internal disturbance set. The novel flexible tube parameterization increases QP feasible region and leads to less conservative results. The switching logic can reduce the conservatism under conventional conditions. Numerical simulation and HIL experiment verify the superiority of the proposed framework. Abstract: Model mismatch caused by strong nonlinearity and other factors will severely impact the lateral path tracking control in Autonomous Vehicles (AVs) under extreme conditions. Previous studies have focused on guaranteeing robust stability under possible uncertainty realizations through Tube-based Robust Model Predictive Control (TRMPC). However, three deficiencies in TRMPC applications are revealed: unknown disturbance set, simple rigid tube, and excessive conservatism. In this paper, a novel scheme named Varying Zonotopic TRMPC with Switching Logic (SVTMPC) is developed to overcome these limitations. Firstly, a zero steady-state error dynamic model is established, and a new update mechanism of the nominal state is devised to determine the unknown internal disturbance set of the AV system. Secondly, zonotopic representation of all defined sets is used to construct the prediction model, as well as a flexible tube with varying cross-sections is naturally designed to overcome excessive conservation and non-solution of Quadratic Programming (QP). Finally, a switching logic between conservative and radical strategies improves tracking performance under conventional conditions without compromising robust stability. Numerical simulation through three scenarios shows that the SVTMPC controller can comprehensively improve robust stability and adaptability compared with MPC and TRMPC. Hardware-in-the-Loop (HIL) experiment verifies the effectiveness and real-time of the SVTMPC controller.