Dynamic path planning and path following control for autonomous vehicle based on the piecewise affine tire model

This paper proposes a novel dynamic path planning and path following control method for collision avoidance, which works based on an improved piecewise affine tire model. The main contribution of this work is the design of a dynamic path planning method based on model predictive control, where it replans a maneuverable path to avoid moving obstacle in real time. A hierarchical control framework contains a high-level path replanning model predictive control and a low-level path following model predictive control. A collision avoidance cost function in the high hierarchies was designed to calculate the relative dynamic distance, which copes with the sudden obstacle. Moreover, the replanning path is the optimized output according to reference trajectory, obstacle, and handling stability. The control objective of the low hierarchies is to accurately track the replanning path, especially for the increased nonlinearity of large tire sideslip angle. For this reason, an improved piecewise affine tire model is designed and used for model predictive control to improve the path following performance and reduce calculated burden. The main improvement of the piecewise affine tire model is that the varied lateral stiffness coefficients adapt to the change of the tire sideslip angle in different tire regions. Based on the CarSim and Simulink platform, the dynamic path planning and path following simulations are designed to test the proposed method. The simulation results demonstrate the effectiveness of the proposed method.