基于隐式PIGPC的网络控制系统时延补偿方法

Implicit PIGPC compensation method for time-delay in networked control system

网络控制系统的时延会导致系统性能下降,针对该问题提出一种基于隐式PIGPC的时延补偿方法。首先,通过历史时延数据,利用自适应变步长最小均方滤波算法对当前时刻的网络时延进行预测,将预测时延代替实际时延作为时延补偿控制器的参数。然后为改进隐式GPC的时延补偿效果,提出具有PI结构的隐式PIGPC算法。隐式PIGPC不必递推求解Diophantine方程,可节省在线计算时间,并且把PI的反馈结构与GPC的预测功能结合起来,改善了控制器的性能。利用隐式PIGPC设计控制器,结合预测时延完成基于隐式PIGPC的网络时延补偿控制器的设计,并对时延补偿算法的稳定性进行了分析。最后的仿真实验表明隐式PIGPC对网络控制系统时延具有良好的补偿效果,同时具有较少的运算时间。

A network time-delay compensation method based on implicit PI-based generalized predictive controller （PIGPC） was presented. The variable step size least mean square algorithm was used for time- delay prediction. The predicted time-delay was used instead of the actual time-delay as the parameters of the network time-delay compensation controller. In order to improve the compensation effect of implicit GPC, an implicit generalized predictive control algorithm with PI structure was used to design the control- ler. Implicit PIGPC doesn＇t need to recursive solve Diophantine function and saves online computing time. The feedback structure of PI and the prediction function of GPC were combined to improve the per- formance of the controller. The time-delay compensation controller was built based on implicit PIGPC and predictive time-delay, and at the same time, the stability of the time-delay compensation algorithm was analyzed. Finally, the simulation results show that the implicit PIGPC has good compensation effect for the time-delay of the networked control system with less computation time.