一类网络控制系统的乱序与量化反馈控制研究

发布时间：2018-01-13 10:28

本文关键词：一类网络控制系统的乱序与量化反馈控制研究　出处：《浙江理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着现代科技在军事、工业、商业领域不断渗透与发展,传统点对点的控制系统已经不能满足人们日益发展的需求,因此需要一种新的控制系统来弥补传统控制系统的不足,进而促进了网络控制系统的发展。网络控制系统是一种将分布在空间中不同位置的系统元件如传感器、控制器、执行器等,通过实时网络连接的反馈控制系统。这种基于网络的控制系统具有成本低、接线少、安装维护方便、可远程操作和控制等优势。但是,由于在控制回路中加入了通信网络,给系统带来了许多新的问题与挑战,如网络诱导时延、数据包丢失与乱序、量化等,这些问题会不同程度地影响系统的控制性能甚至破环系统的稳定性。目前,对网络控制系统课题的研究已成为了国内外控制领域的一个研究热点。本文主要从系统的建模、稳定性分析、控制器设计三个方面分别对网络控制系统中的数据包乱序问题和量化问题进行了研究,主要工作总结如下:第一:针对一类网络控制系统中出现的数据包乱序问题,提出了一种补偿控制方法以消除数据包乱序对系统的不利影响。首先利用时间戳(time-stamp)方法对从源节点发送的数据包时间序列进行标记,再在控制器和执行器端分别设置一个比较器,两者结合来判断数据包是否发生乱序,若乱序则对数据包做相应处理,这使得用于被控对象的控制信号总保持最新,闭环系统被建模成一个Markov跳变系统。运用Lyapunov稳定性理论得到了系统随机稳定的充要条件,这个条件用一组具有逆约束的线性矩阵不等式(LMIs)表示。将控制器设计问题转化为线性矩阵不等式的凸优化问题,利用锥补线性化算法(CCL)得到了状态反馈控制器的增益。第二:讨论了网络诱导时延和信号量化对网络控制系统的影响,基于对数量化反馈控制器和零阶保持器的工作机制,利用扇形界方法,将系统建模成带有输入时延的不确定时滞系统。其中,在系统的反馈通道和前馈通道分别加入了对数量化器,采用时延依赖的Lyapunov-Krasovskii泛函方法,得到了系统的稳定性判据,这个稳定性条件是以两个线性矩阵不等式(LMIs)给出的。设计量化反馈控制器使得系统渐近稳定,并指出量化器的量化密度直接影响系统的控制性能。最后给出了数值仿真示例验证了此方法的有效性。
[Abstract]:With the development of modern science and technology in the military, industrial, commercial areas continue to penetrate and develop the traditional point-to-point control systems have been unable to meet the growing demand of development, so we need a new control system to make up for the deficiencies of the traditional control system, and promote the development of the network control system. Networked control system is a kind of distribution in the space position of different system components such as sensors, controllers, actuators, through real-time network connection of the feedback control system. This control system based on network has the advantages of low cost, less wiring, convenient installation and maintenance, remote operation and control. However, due to the addition of the communication network in the control loop, bring many new problems and challenges to the system, such as network induced delay, packet loss and disorder, quantification, these problems will affect the control performance of the system even The stability of broken ring system. At present, the research on control system of project network has become a research hotspot in control field. This paper mainly from the stability analysis of system modeling, controller design, three respectively on the network control system of packet reordering problem and the quantization problem is studied, the main summary of the work are as follows: first, aiming at the problem of disorder there is a kind of network control system in data packets, a compensation control method to eliminate the adverse effect of packet reordering on the system. Firstly, the timestamp (time-stamp) method for packet time from the source node sends the data sequence tag, then set a comparator in the controller and the actuator side, both to determine whether the packet reordering occurs, if the order of packets corresponding treatment, which makes the total control signal for a controlled object Keep up-to-date, the closed-loop system is modeled as a Markov jump system. Using the Lyapunov stability theory, sufficient and necessary conditions for the stochastic stability of system is obtained. This condition with a set of linear matrix inequalities with inverse constraints (LMIs). The controller design problem is transformed into a convex optimization problem with linear matrix inequalities, using cone complementary linearization algorithm (CCL) to get the state feedback controller gain. Second: discuss the network induced time delay and signal quantization effects on the network control system, quantization feedback controller and the zero order holder working mechanism based on using the sector bound approach, the system is modeled as uncertain time-delay systems with input delay. And the logarithmic quantizer were added in the system of the feedback channel and the feedforward channel, using Lyapunov-Krasovskii functional method, delay dependent stability criterion, this system has been. A stability condition is based on two linear matrix inequalities (LMIs) are presented. The design of quantized feedback controller such that the asymptotic stability of the system, and points out that the quantizer quantization density directly affects the control performance of the system. Finally, numerical simulation results verify the effectiveness of this method.

【学位授予单位】：浙江理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP273

【参考文献】