Abstract:

The multi-axle vehicle dynamics model was established based on the valve-control cylinder of 4 axis (8×2) heavy duty truck. Threshold control and PID online setting based on RBF(radial basis function) neural network control algorithm were used to optimize steering stability and steering wheel-wear. After simulating the dynamics model in MATLAB/Simulink, the time domain response of yaw velocity, centroid side-slip angle and lateral acceleration to the first axle angle step under the conditions of 20km/h and 50km/h were obtained and converted into bode diagram with frequency domain analyses. At the same time, under the conditions of running steering and parking steering, 3 kinds of corresponding step time domain response of the third axle angle with the open loop control, PID control and RBF neural network setting PID control were compared. The simulation results show that the vehicle has better operation stability and tire abrasion resistance. 

Key words: electronically controlled hydraulic system, neural network PID control, operation stability, steering wheel tire abrasion

摘要：

针对某8×2重型货车，建立了阀控缸的二自由度多轴车辆转向动力学模型。使用门限阈值控制及基于RBF神经网络的PID在线整定控制方法，以操纵稳定性和轮胎磨损最优为控制目标，在MATLAB/Simulink中进行仿真，得到了20km/h和50km/h时横摆角速度、质心侧偏角、侧向加速度在第一轴转角阶跃输入下的时域响应，并进行了bode图的频域分析。同时对比了开环控制、PID控制、RBF神经网络整定PID控制下，第三轴转角在行驶及原地转向时的阶跃时域响应。仿真结果表明，车辆具有较好的操纵稳定性和轮胎抗磨性。

关键词: 电控液压系统, 神经网络PID控制, 操纵稳定性, 转向轮胎磨损