一类有数据包丢失的长时延网络控制系统的最优控制  

Optimal Control for A Class of Long Delay Networked Control System with Data Packet Dropout

作　　者：李海涛[1] 唐功友[1] 

机构地区：[1]中国海洋大学信息科学与工程学院,山东青岛266100

出　　处：《中国海洋大学学报（自然科学版）》2010年第7期131-135,共5页Periodical of Ocean University of China

基　　金：国家自然科学基金项目(60574023;40776051);山东省自然科学重点基金项目(Z2005G01)资助

摘　　要：研究一类具有数据包丢失的长时延网络控制系统的最优控制问题。首先建立控制器和执行器均采用时钟驱动的长时延网络控制系统的离散时间系统模型,提出1种基于维序策略的数据包丢失判别方法。然后在数据包的最大丢包率和最大连续丢包长度被限定的情况下,提出1种用于补偿丢失数据的估值补偿方法。该方法可以根据历史状态向量和控制向量信息估算出当前的控制量。利用该方法和动态规划算法,导出了离散二次型性能指标下的网络控制系统的最优控制序列,进而给出了最优控制律的实现算法。最后通过1个数值仿真算例,比较了不对丢失数据进行补偿和采用估值补偿2种方法对于系统性能的影响,从而验证了主要结果的有效性和正确性。For the long delay Networked Control Systems with data packet dropout, the problem of optimal control is researched. Firstly, while controller and actuator are clock-driven, a discrete-time model is established, and a judgement method of data packet dropout based on maintains sequence strategy is presented. Then, under the situation that the maximum probability and successive length of data packet dropout are restricted, a data packet dropout value-estimated compensating method is presented. This method can evaluate the current control vector according to historical state vector and control vector information. By using the method and dynamic programming algorithm, the optimal control sequence that based on discrete quadratic criteria is derived; furthermore, an implementation algorithm of the optimal control law is given. Finally, a numeric example compares the value-estimated compensating method with the method of no value-estimated compensating when data packet dropouts happen. The results illustrate the effectiveness and the correctness of the main conclusion.

关 键 词：网络控制系统 最优控制 数据包丢失 估值补偿 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程] TP273[自动化与计算机技术—控制科学与工程]