一种新的基于二阶神经网络的执行器死区补偿方法

Dead-zone Compensation Method for Actuator Based on Second-order Neural Network

针对执行器的未知宽度的死区非线性环节,提出一种基于神经网络死区补偿的新方法。该补偿器采用了两个神经网络。一个是用来预测未知的死区,另外一个则在前向反馈通路中提供适当的补偿。仿真结果表明该补偿策略对于对称的和非对称的死区非线性都是有效的。

A new method about dead-zone compensation based on second-order neural network is presented to solve the unknown width dead-zone nonlinear in actuator. The compensator uses two neural networks, among them, one is used to estimate the unknown dead-zone, the other is used to provide adequate compensation in the feed-forward path. Simulation shows that this compensation strategy is effective for both symmetrical and nonsymmetrical dead-zone.