Abstract: Aiming at the problem that the control precision is low when the traditional artificial potential field method is applied to the lane keeping task, a new lane keeping assist control method based on parameter varying artificial potential field is proposed. Firstly, considering the effect of vehicle states on lane keeping control, two variable parameters, longitudinal velocity and lateral velocity, are used to construct the road potential function, and the desired steering angle is solved by the potential field gradient. The design parameters of the potential field function are determined by the trajectory prediction theory. Then the state equation model with respect to the path tracking error containing the variable parameters is established and the stability of the closed-loop control system is analyzed. On this basis, an assist torque controller is designed to achieve lane keeping. To take into account the human-machine coordination, the driver torque and driver intention path are used to design fuzzy rules to dynamically adjust the assist control weight. Simulation results show that the proposed method can effectively improve the lane control precision and achieve a better human-machine coordination performance. The method is also verified on the CarSim/LabView hardware-in-the-loop test bench. The conclusions are in accordance with the simulation.

Key words: artificial potential field, coordination control, fuzzy rule, lane keeping assist, parameter varying system