恒压网络静液传动系统的神经网络滑模控制

根据恒压网络条件下的静液传动系统的特点,建立用于转速控制的二自由度动力学模型.针对恒压网络静液传动系统的参数摄动和不确定性,选择液压泵/马达的角速度和角加速度为控制变量,设计一种神经网络自适应滑模控制器,采用径向基函数神经网络(RBFN)取代滑模切换控制部分,利用其在线学习功能,对系统的不确定因素进行自适应补偿,应用李亚普诺夫稳定性理论推导网络权值的在线自适应率,保证闭环控制系统的稳定性.在模拟试验台上进行了阶跃信号和斜坡信号的转速控制响应分析,并与常规PID控制以及基于神经网络的PID(NNPID)控制进行对比.试验结果表明:所设计的控制器具有良好的控制效果,能使系统具有良好的跟踪性和强的鲁棒性,有效地消除高频抖振现象.

Neural network sliding mode control for constant pressure hydrostatic transmission system

A two degrees of freedom vehicle dynamic model was set up for speed control according to the characteristics of constant pressure hydrostatic transmission system.Then selecting angular velocity and angular acceleration of hydraulic pump/motor as the control variables,a novel neural network sliding mode control strategy was proposed,which was applied to ensure tracking capability to constant pressure hydrostatic transmission system in the presence of plant parameter variations and uncertainties.A radial basis function neural network(RBFNN) was utilized to realize the corrective control of sliding mode control,and compensate uncertainties of the system with adaptive learning algorithm,the parameter on-line adaptive laws were derived in the sense of Lyapunov stability theorem to guarantee the system stability.The speed control and response analysis of step signal and ramp signal were conducted in the simulation test platform,and contrasted with the conventional PID control and the PID based on the neural network(NNPID) control.The experimental results show that the proposed control scheme has good tracking performance and strong robustness,and eliminates chattering effectively.