【作者】 陈刚；

【导师】 阳春华；

【作者基本信息】 中南大学 ， 计算机应用技术， 2012， 博士

【摘要】 网络控制系统是近年来国际控制领域研究的热点问题之一。网络控制系统的重要特点是其用于信息传输的通信网络具有不确定性。有限的网络带宽使得信息和数据在通信网络中传输出现时滞、数据包丢失等问题,这些问题的出现导致网络控制系统性能下降和不稳定。如何合理地设计网络控制系统,确保整个网络控制系统的稳定性以及控制性能,是网络控制系统的重要研究内容之一。本文针对存在网络诱导时滞和丢包问题的网络控制系统,建立了具有网络诱导时滞和丢包问题的线性与非线性网络控制系统新模型；提出了具有更低保守性的网络控制系统镇定控制器与鲁棒H∞滤波器设计方法。论文的研究工作和创新性成果包括：(1)通过引入具有最新信号选择功能的逻辑零阶保持器(ZOH),建立了一种新的将网络诱导时滞和丢包的影响统一转化为时滞问题的网络控制系统模型。利用Lyapunov-Krasovskii泛函并结合积分不等式方法,提出了一种改进的线性网络控制系统稳定性分析方法。并基于改进的锥补线性化迭代算法(ICCL)给出了一种具有更低保守性的控制器设计方法。数值仿真表明所提出方法具有更低的保守性。(2)针对同时具有网络诱导时滞和丢包问题的非线性Lurie网络控制系统,提出了一种新的Lurie网络控制系统的时变时滞模型。在构造Lyapunov-Krasovskii泛函时包含网络诱导时滞的上下界,利用积分不等式方法,提出了一种非线性Lurie网络控制系统的时滞状态反馈控制器的设计方法。数值仿真表明了该方法的有效性和优越性。(3)针对基于T-S模糊模型描述的非线性网络控制系统,利用输入时滞法和并行分布补偿(PDC)控制原理,建立了一种新的T-S模糊网络控制系统的时滞模型。然后,利用Lyapunov-Krasovskii泛函并结合自由权矩阵方法,不忽略任何Lyapunov-Krasovskii泛函过程中的有用项,提出了一种改进的基于LMI的T-S模糊网络控制系统的稳定性分析方法。通过求解一组LMI,获得模糊控制器增益矩阵。仿真实例表明该方法得到的时滞上界具有更低的保守性。(4)针对一类具有时变网络诱导时滞和丢包问题的Lurie网络控制系统,建立了一种新的Lurie网络控制系统的多时滞模型。通过保留泛函导数中经常被忽略的有用项,提出了一种新的多时滞Lurie非线性网络控制系统的稳定性分析方法。在控制器设计方面,通过增加一等式约束将NLMI转化为LMI表示的可解性问题,提出了一种新的具有更低保守性的动态输出反馈控制器设计方法。通过理论证明和仿真例子验证了所提方法的优越性。(5)研究了网络控制系统中的鲁棒H∞滤波问题。考虑到系统的测量信号通过具有网络诱导时滞和丢包问题的网络传输到滤波器,分别建立了基于连续时间与离散时间的网络控制系统滤波器的模型,采用Lyapunov-Krasovskii泛函和积分不等式方法,通过对系统中的不确定性问题进行处理,建立了一种新的鲁棒H∞滤波器误差系统的鲁棒H∞性能分析准则。基于H∞性能分析的结果,得到了满足条件的H∞滤波器存在的充分条件,同时确定了滤波器的系数矩阵。数值实例表明该设计方法具有更低的保守性。更多还原

【Abstract】 Networked control systems is one of the hot issues of the control field in recent years. The important feature of the networked control systems is that the communication network for information transmission is uncertainty. Limited network bandwidth makes information and data appear delay and packet loss in communications. These problems lead to networked control systems performance degradation and instability. How to design rationally networked control systems to ensure the stability and control performance of the entire networked control systems is one of the important research issues on the networked control systems.In this dissertation, based on the networked control systems existing network-induced delay and packet loss, we have established a new model of the linear and nonlinear networked control systems with network induced delay and packet loss issues. We have also proposed some lower conservative design methods for stabilization controller and robust H∞filter in networked control systems. Main results of this dissertation are as follows:(1) By introducing a logical zero-order hold (ZOH) with the latest signal selection function, we have established a new linear networked control systems model with account of network-induced delay and packet loss. Using Lyapunov-Krasovskii function combined with integral inequality, we have developed an improved stability analysis method for the linear networked control systems, and further based on an improved cone complementarity linearization iterative algorithm(ICCL), a lower conservative controller design method has been presented. Numerical simulations show that the proposed method has a lower conservative.(2) By taking into account non-linear Lurie networked control systems with network-induced delay and packet loss issues, we have put forward a new kind of time-varying delay model for Lurie networked control systems. We have also developed a new design method for nonlinear Lurie networked control systems with time-delay state feedback controller by constructing Lyapunov-Krasovskii function with upper and lower bounds of the network-induced delay, and through integral inequality approach. Numerical results show that the method is effective and advantageous.(3) For the nonlinear networked control systems with T-S fuzzy model, using the input-delay and parallel distributed compensation(PDC) techniques, a new model has firstly been established in which both network-induced delay and data loss are taken into consideration, and then using Lyapunov-Krasovskii function and combined the free-weighting matrix method, we have developed an improved networked control systems stability analysis method based on LMI T-S model without overlooking the useful items in the Lyapunov-Krasovskii functional derivative. By solving a set of LMI, fuzzy controller gain matrix is obtained. The simulation results show that the delay upper bound based on this method has lower conservative.(4) For the Lurie networked control systems with time-varying network-induced delay and packet loss, we have developed a new multi-delay Lurie nonlinear networked control systems model. Furthermore, through retaining useful items often overlooked in the Lyapunov functional derivative, we have developed a multi-delay Lurie nonlinear networked control systems stability analysis method. In the controller design, we have proposed a new dynamic output feedback controller design method through the transformation of the non-linear LMI into the linear LMI solvability problem. By theoretical analysis and simulation examples, the advantages of the proposed method are verified.(5) Study on robust H∞filter issues in the networked control systems. Considering the measured signals is transmitted to the filter through the system, being inevitably subjected to the effect of the network-induced delay and packet loss, we have established networked control systems based on continuous-time and discrete-time model. By using Lyapunov-Krasovskii functional and integral inequality method to deal with the uncertainty in the system, we have established a new kind of robust H∞performance analysis criteria for networked control systems. Based on H∞performance analysis results, we have achieved the sufficient conditions for the existence of the H∞filter and determined the filter coefficient matrix. Numerical examples demonstrate that the design method is less conservative.更多还原