Vehicle Lane Keeping Single Neural Network Control with Driver Behavior

According to the nonlinear and parameter time-varying characteristics of vehicle lateral dynamics, a novel algorithm of vehicle lateral adaptive PID control with single neuron network was proposed. Based on self-learning and adaptive ability of single neural network, the parameters of vehicle lateral PID controller were self-tuning on-line and the problem of large computation time brought by traditional adaptive PID control was avoided, in which the parameters of reference model of the controlled system must be identified with large calculation burden. The results of the simulation and field experiment show this algorithm can effectively control vehicle to keep and track the pre-given target trajectory, and has good robustness and adaptability for vehicle lateral nonlinear dynamics.