Active Disturbance Rejection Control of Certain Balanced and Positioning Electro-Hydraulic Servo System Based on Neural Network

Abstract

For certain balance and positioning of electro-hydraulic servo system existing nonlinear, the active disturbance rejection control (ADRC) has many adjustable parameters which are difficult to regulate, so active disturbance rejection control with neural network (NN-ADRC) is developed in this paper. The method uses neural network self-learning ability, through a single neuron adaptive configuration parameters to complete an online self-tuning parameters, while taking advantage of RBF neural network as identifier to identify the controlled object gradient information. Simulation results show that: the controller parameters are reduced significantly, and effectively inhibit the system unbalance force disturbance and realize the accurate positioning. It also has fast response speed, no overshoot, and high steady-state accuracy.