网络控制系统随机延迟的预测控制  

Predictive Control Of Networked Control Systems Random Delay

作　　者：陈在平[1,2] 王雪[1,2] 

机构地区：[1]天津理工大学天津市复杂系统控制理论及应用重点实验室,天津300384 [2]天津理工大学自动化学院,天津300384

出　　处：《计算机仿真》2014年第9期327-330,共4页Computer Simulation

基　　金：天津市科技支撑重点项目(10ZCKFGX03000);科技部中小企业创新基金(09C262112022954);天津市自然科学基金重点项目(07JCZDJC09600);天津高校科技发展基金(20120705)

摘　　要：针对网络控制系统随机时延、时变等特性造成控制效果不理想的问题,提出一种改进的广义预测控制与递推最小二乘法相结合的预测控制策略。上述控制策略将传统的预测控制运用于网络控制系统中,针对网络随机时延的特性,采用改进的广义预测控制算法对随机时延进行补偿。结合分段可变遗忘因子递推最小二乘法对被控对象模型参数进行在线辨识,可以减少网络控制系统参数突变对系统造成的不利影响。仿真结果表明,所提出的预测控制策略可以有效解决网络控制系统随机时延问题,并且在参数突变的情况下也能够获得良好的控制效果。In order to solve the problem of the unsatisfied control performance caused by random delay and timevarying property in Networked Control Systems （NCSs） , a control strategy based on improved predictive general control and recursive least square is proposed. In this control strategy, according to the characteristics of network random time-delay, an improved generalized predictive control （GPC） algorithm is adopted to compensate the random time- delay. In addition, combining the recursive least squares with a variable forgetting factor algorithm to indentify the model parameters of controlled object on-line in different stage, it can reduce the adverse effect on system due to pa- rameters changes of NCSs. Simulation results show that the improved predictive control strategy proposed can deal with the random time-delay in NCSs effectively, and also get a well control resuh under the terrible condition of parameters change.

关 键 词：网络控制系统 广义预测控制 随机时延 参数突变 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]