Abstract: Three axle independent steering vehicles are widely used in special fields and are gradually developing towards unmanned and intelligent, with path tracking control methods being an important research content. Three axle independent steering vehicles have more complex steering behaviour compared to traditional vehicles, so it is necessary to fully consider the impact of their steering characteristics in the path tracking control process. This study focuses on the steering control problem of a three axis independent steering unmanned vehicle during path tracking. A multi degree of freedom model is established based on theoretical mechanics and tire magic formulas, and the differences in lateral displacement and yaw angle changes under different steering modes are compared. Then, using the actual vehicle response data as a reference, the non dominated sorting genetic algorithm II(NSGA-II) was used to optimize the multi degree of freedom model parameters, and the dataset was obtained through model simulation. The steering mode switching strategy is developed through the BP neural network, and finally the mode switching strategy is introduced into the model predictive controller. Simulation and actual vehicle test results show that the tracking accuracy of three-axis independent steering vehicles can be effectively improved through steering mode switching during the path tracking process.

Key words: three-axis vehicle, independent steering, mode switch, neural network, path tracking