基于前馈神经网络复合逆控制的研究

Study on complex inverse control based on feedforward neural network

为了提高神经网络直接逆控制方法的跟踪精度和抗干扰能力,结合前馈神经网络直接逆控制与通用模型控制策略各自的优点,提出了一种基于前馈神经网络复合逆控制的方法。该方法将控制系统的参考轨迹改造成一条规范的二阶曲线,从而使得控制器参数物理意义明确,且易于整定。仿真实验验证了系统具有良好的鲁棒性和抗扰性能。

A complex inverse control strategy was developed based on the advantages of feedforward neural network direct in- verse control and generic model control to improve the tracking precision and disturbance rejection of neural network direct inverse control methods. The strategy transforms reference trajectory of the control system into a classic second- order curve. The parame- ters of this controller have explicit physics meaning and it is very easy to tune. The results show that the system has strong robust- ness to the variation of system parameters and the disturbance of load.