节点文献

具有时延及丢包的不确定网络控制系统的模型预测控制

Model Predictive Control of Uncertain Network Control Systems with Time-delay and Data Packet Loss

【作者】 郑伟

【导师】 陈东彦;

【作者基本信息】 哈尔滨理工大学 , 应用数学, 2012, 硕士

【摘要】 网络控制系统中的传输时延和数据丢包是不可避免的,时延和丢包常常会导致系统不稳定.利用模型预测控制方法设计对网络时延和数据丢包具有一定补偿作用的网络模型预测控制,可改善系统的稳定性.另外,在对实际系统建模时,由于内部机理和外部环境的影响,使得所建模型不能精确地描述实际问题,实际系统模型中必会存在着不确定性,而不确定性也是导致系统不稳定的主要原因.因此,在网络环境下研究不确定系统的控制问题具有重要的理论意义和实际价值.本文研究具有时延和丢包的不确定网络控制系统的模型预测控制方法,主要内容如下:1.研究具有时延的不确定网络控制系统的模型预测控制策略.假设在传感器与控制器、控制器与执行器两端都通过网络连接,且网络数据存在传输时延.在系统状态不可测的情况下,设计系统的状态观测器;利用min-max鲁棒模型预测控制方法设计预测控制器;利用鲁棒模型预测控制的控制冗余,设计网络时延补偿器,以解决网络时延对系统稳定性的影响;设计网络控制系统模型预测控制算法,实现对系统的稳定控制.2.研究具有数据丢包的不确定网络控制系统的模型预测控制方法.假设在传感器与控制器、控制器与执行器通道均具有随机数据丢包.利用未丢失数据在线估计丢失数据,适当选取预测时域和优化界限,运用min-max鲁棒模型预测控制方法设计鲁棒模型预测控制器,并设计相应的网络控制系统模型预测控制算法.

【Abstract】 Transmission delay and data packet loss in the network control system areinevitable, and they will often lead to instability of the system. By modelpredictive control method and the design of network model predictive control tocompensate the network time delay and data packet loss, we can improve thestability of the system. In addition, in the practical system modeling, because ofthe influence of the internal mechanism and the external environment, the modelcan’t accurately describe the practical problem, so there are uncertainties in thepractical system model. Uncertainties are the main reason of instability.Therefore, the study of the problems of uncertain control system under thenetwork environment will have important theoretical and practical value. In thisthesis, the model predictive control method is invest igated in the network contro lsystem with time delay and data packet loss and the main contents are organizedas follows:1. The model predictive control strategy of the uncertain network controlsystem with time delay is studied. Assume that sensor and controller, as well ascontroller and actuators, are connected with network and that there istransmission delay in the network. A state observer is designed under the systemstates unobservable. Min-max robust model predictive control method is used todesign a predictive controller. The control redundancies of robust modelpredictive control are used to design the network time delay compensator, inorder to solve the influence of time delay on the stability in the network system.A model predictive control algorithm of the stability control o f the system isachieved.2. The model predictive control method of uncertain network control system with data packet loss is studied. Assume that there is random data packets loss inthe channels between sensor and controller, and between controller and actuators.By using data which is not lost to estimate online the lost data, and selectingappropriate prediction horizon and optimizing boundaries, based on min-maxrobust model predictive control method a robust model predictive controller isdesigned, and the corresponding model predictive control algorithm of networkcontrol system is designed.

节点文献中: 

本文链接的文献网络图示:

本文的引文网络