Holistic Adaptive Multi-Model Predictive Control for the Path Following of 4WID Autonomous Vehicles

Yixiao Liang,Yinong Li,Amir Khajepour,Ling Zheng
DOI: https://doi.org/10.1109/tvt.2020.3046052
IF: 6.8
2020-01-01
IEEE Transactions on Vehicular Technology
Abstract:This paper presents a novel adaptive multiple-model predictive control (MPC) scheme for Four-Wheel Independent Drive (4WID) autonomous vehicles with a holistic structure. Firstly, the combined vehicle-path model is established. To ensure the real-time performance of MPC, the coupling relationships in the control output and the longitudinal/lateral motions are well decoupled, in this way a linear integrated model is utilized as the internal model of the controller. Then, the holistic MPC is proposed to acquire the steering angle and the force on each corner. Based on the advantages of the proposed structure, a weight adaptive mechanism is introduced to improve the handling ability of the controller to various driving conditions, especially some extreme conditions. For the uncertainties in tire cornering stiffness, a multiple-model adaptive law is designed with its convergence proved by Lyapunov theory. Numerical results based on Carsim-Simulink co-simulation platform demonstrate the effectiveness and superiority of the proposed control method in both normal and extreme driving conditions.
What problem does this paper attempt to address?