Steering neural network PID control for tracked vehicle with hydrostatic drive

Based on steering kinematics and dynamics analysis of tracked vehicle, steering control strategy was presented to realize reducing average vehicle speed automatically while achieving the driver's expected steering radius exactly in the case of not exceeding the system pressure threshold and secure steering. The steering controller was comprised of neural network PID controller and pump & motor displacement controller. The steering neural network control simulation was conducted by using Simulink of Matlab. The simulation results indicated that compared with conventional PID control, neural network control export overshoot reduced from 10.5% to 4.1% and control response time decreased from 4.8 s to 2.2 s, which meant that system real-time ability and robustness were improved. The simulation results for various steering conditions demonstrated that good steering stability and maneuverability were obtained with neural network control for tracked vehicle with hydrostatic drive.