基于混杂系统理论的网络化控制分析与设计
发布时间：2019-06-02 01:01
【摘要】：随着计算机网络技术的快速发展,以网络为传输介质的网络化控制系统已成为控制系统的一个很重要的发展方向。网络化控制系统是一种基于网络的分布式控制系统,与传统点到点的控制系统相比,具有共享信息方便、布线少、易于系统维护和扩展等优点。然而,网络的引入也给控制系统的分析和设计带来了新的挑战。在网络化控制系统领域,虽然目前已经取得了一些重要的研究成果,但是仍然存在很多问题亟待解决。本文主要是运用异步动态系统和切换系统等混杂系统理论分别从考虑调制的网络化控制系统、带有介质接入约束的网络化控制系统和基于模型的网络化控制系统等方面开展研究的,取得的研究成果如下：(1)针对考虑调制的网络化控制系统,提出了将误码率考虑为丢包率的策略,填补了目前网络化控制系统领域处理调制信号错码问题的空白。文中分别从二进制调制和多进制调制的角度进行了研究：对于较为简单的二进制调制,将带有常时延和错码的网络化控制系统建模为一个异步动态系统,借助Lyapunov稳定性理论进行了稳定性分析和状态反馈控制器参数的设计,并得到了使带有二进制调制的网络化控制系统稳定的误码率定理；对于较为实用却更加复杂的多进制调制,将带有常时延和错码的网络化控制系统建模为一个异步动态系统,借助Lyapunov理论进行了稳定性分析和动态输出反馈控制器的设计,得到了使带有多进制调制的网络化控制系统稳定的误码率定理,并给出了多进制调制的进制选择定理。(2)针对带有介质接入约束的网络化控制系统,首次提出了控制与调度耦合设计的概念,并在分析现有文献中基于通信序列和最大误差优先-尝试一次丢弃的控制与调度耦合设计方法的基础上,提出了基于系统理想动态的类最小二乘调度策略。在将不含时延的系统的理想动态嵌入到调度器后,在每个采样周期内,将当前状态与理想动态的状态进行比较,具有较大偏差的状态优先获得接入网络的权限,有效地降低了系统分析和设计的保守性。在所提出的类最小二乘调度算法的基础上,将含有随机短时延和介质接入约束的网络化控制系统建模为一个切换系统,借助切换系统理论和Lyapunov稳定性理论进行了类最小二乘调度与控制的耦合设计。此外,进一步改进了所提出的类最小二乘调度算法,将含有时延的理想动态代替了不含时延的理想动态,因含有时延的理想动态能更全面地反映真实的系统,因而改进后的调度策略具有更好的性能。(3)针对用于节省带宽的基于模型的网络化控制系统方法,提出了用切换系统的平均驻留时间方法来进行分析和设计的策略,提高了系统分析和设计的灵活性。通过将基于模型的网络化控制系统建模为切换系统,利用Lyapunov稳定性理论和切换控制的平均驻留时间方法对基于模型的网络化控制系统进行了指数稳定性分析和状态反馈控制器的设计。此外,对现有文献中带有观测器的基于模型的网络化控制系统架构进行了改进,将观测器由接入网络前移到控制器中,这种改进不仅节省了设计成本还避免了因网络丢包所造成的信号丢失的问题。在将带有观测器的基于模型的网络化控制系统建模成切换系统后,运用Lyapunov稳定性理论和切换控制的平均驻留时间方法对基于模型的网络化控制系统进行了指数稳定性分析和控制器与观测器的设计。(4)针对基于模型的网络化控制系统,提出了用脉冲切换控制的平均驻留时间方法来进行分析和设计的策略,降低了系统分析和设计的保守性。通过分析网络闭合时系统状态存在跃变的现象,将基于模型的网络化控制系统建模为脉冲切换系统,利用Lyapunov稳定性理论和脉冲切换控制的平均驻留时间方法对基于模型的网络化控制系统进行了指数稳定性分析和状态反馈控制器的设计。此外,在改进的带有观测器的基于模型的网络化控制系统的基础上,将带有观测器的基于模型的网络化控制系统建模成脉冲切换系统,运用Lyapunov稳定性理论和脉冲切换控制的平均驻留时间方法对基于模型的网络化控制系统进行了指数稳定性分析和控制器与观测器的设计。(5)针对基于模型的网络化控制系统,提出了脉冲切换控制的依赖于模式的平均驻留时间策略,进一步降低了系统分析和设计的保守性。在脉冲切换控制的依赖于模式的平均驻留时间策略中,系统的平均驻留时间不仅依赖于系统的切换参数,还与系统的脉冲矩阵参数有关。通过在脉冲切换控制的依赖于模式的平均驻留时间策略下对基于模型的网络化控制系统进行分析和设计,得到了使得系统指数稳定的充分条件和状态反馈控制器。此外,利用脉冲切换控制的依赖于模式的平均驻留时间策略对带有观测器的基于模型的网络化控制系统进行了分析和设计,给出了使得系统指数稳定的充分条件和控制器与观测器的设计。论文通过Matlab和近似网络环境的TrueTime工具箱对一些实例进行了数值仿真,验证了文中所提出的方法的有效性。最后,对全文的工作进行了总结,并对需要进一步研究的工作提出了建议。
[Abstract]:With the rapid development of the computer network technology, the networked control system with the network as the transmission medium has become an important development direction of the control system. The networked control system is a kind of distributed control system based on the network. Compared with the traditional point-to-point control system, it has the advantages of convenient sharing information, less wiring, easy system maintenance and extension. However, the introduction of the network presents a new challenge to the analysis and design of the control system. In the field of networked control system, although some important research results have been achieved at present, there are still many problems to be solved. This paper mainly uses the hybrid system theory such as the asynchronous dynamic system and the switching system to carry out the research from the networked control system, the networked control system with the medium access constraint and the network control system based on the model, respectively, and the results are as follows: (1) Aiming at the networked control system considering the modulation, the strategy of considering the bit error rate as the packet loss rate is put forward, and the blank of the problem of processing the modulation signal detection code in the field of the present networked control system is filled. In this paper, the angle of binary modulation and multi-system modulation is studied. For simpler binary modulation, the networked control system with the constant time delay and the error code is modeled as an asynchronous dynamic system. The stability analysis and the design of the state feedback controller parameters are carried out by means of the Lyapunov stability theory, and the bit error rate theorem of the networked control system with binary modulation is obtained; for the more practical and more complex multi-system modulation, The networked control system with the constant time delay and the error detection code is modeled as an asynchronous dynamic system. The stability analysis and the design of the dynamic output feedback controller are carried out by means of the Lyapunov theory, and the bit error rate theorem of the networked control system with the multi-system modulation is obtained. And the system selection theorem of the multi-system modulation is given. (2) For a networked control system with medium access constraints, the concept of control and scheduling coupling design is proposed for the first time, and based on the analysis of the control and scheduling coupling design method based on communication sequence and maximum error in the existing literature, A class least-square scheduling strategy based on the ideal dynamic of the system is proposed. After the ideal dynamic embedding of the system with no time delay is embedded in the scheduler, the state of the current state and the ideal dynamic state is compared with the state of the ideal dynamic state in each sampling period, the state with the large deviation preferentially obtains the authority of the access network, and the system analysis and the design conservatism are effectively reduced. On the basis of the proposed class least square scheduling algorithm, a networked control system with random short-time delay and medium access constraints is modeled as a switching system, and the coupling design of the class least-square scheduling and control is carried out by means of the theory of switching system and the Lyapunov stability theory. In addition, the proposed least-squares scheduling algorithm is further improved, and the ideal dynamics with time delay is replaced by the ideal dynamics without delay, and the real system can be more fully reflected due to the ideal dynamics with time delay, so the improved scheduling strategy has better performance. (3) Based on the model-based networked control system for saving the bandwidth, the method of the average residence time of the switching system is proposed to analyze and design the strategy, and the flexibility of the system analysis and design is improved. The model-based networked control system is modeled as a switching system, and the exponential stability analysis and the state feedback controller design are carried out on the model-based networked control system by means of the average residence time method of the Lyapunov stability theory and the switching control. In addition, the model-based networked control system architecture with an observer in the prior art is improved, and the observer is advanced to the controller by the access network, and the improvement not only saves the design cost but also avoids the problem of signal loss caused by network packet loss. After the model-based networked control system with the observer is modeled as a switching system, the exponential stability analysis of the model-based networked control system and the design of the controller and the observer are carried out using the Lyapunov stability theory and the mean residence time method of the switching control. (4) According to the model-based networked control system, the method of average residence time for pulse-switching control is proposed to analyze and design the strategy, and the conservative property of system analysis and design is reduced. The model-based networked control system is modeled as a pulse switching system by analyzing the phenomena of the system state transitions when the network is closed. The exponential stability analysis of the model-based networked control system and the design of the state feedback controller are carried out using the Lyapunov stability theory and the mean dwell time method of the pulse-switching control. In addition, on the basis of an improved model-based networked control system with an observer, a model-based networked control system with an observer is modeled as a pulse switching system, The exponential stability analysis of the model-based networked control system and the design of the controller and the observer are carried out using the Lyapunov stability theory and the mean residence time method of the pulse-switching control. (5) In view of the model-based networked control system, the model-based average residence time strategy for pulse switching control is proposed, and the conservative property of system analysis and design is further reduced. In the mode-dependent average dwell time strategy of the pulse-switching control, the average dwell time of the system depends not only on the switching parameters of the system but also on the pulse matrix parameters of the system. By analyzing and designing the model-based networked control system under the mode-dependent average residence time strategy of the pulse-switching control, a sufficient condition and a state feedback controller for the stability of the system index are obtained. In addition, the model-based networked control system with the observer is analyzed and designed by means of the average residence time strategy of the mode based on the pulse switching control, and the sufficient conditions for the stability of the system index and the design of the controller and the observer are given. In this paper, some examples are simulated by using the TrueTime toolbox of Matlab and the approximate network environment, and the validity of the proposed method is verified. Finally, the paper summarizes the work of the full text, and puts forward some suggestions on the work in need of further research.
【学位授予单位】：北京交通大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP273
，
本文编号：2490711
[Abstract]:With the rapid development of the computer network technology, the networked control system with the network as the transmission medium has become an important development direction of the control system. The networked control system is a kind of distributed control system based on the network. Compared with the traditional point-to-point control system, it has the advantages of convenient sharing information, less wiring, easy system maintenance and extension. However, the introduction of the network presents a new challenge to the analysis and design of the control system. In the field of networked control system, although some important research results have been achieved at present, there are still many problems to be solved. This paper mainly uses the hybrid system theory such as the asynchronous dynamic system and the switching system to carry out the research from the networked control system, the networked control system with the medium access constraint and the network control system based on the model, respectively, and the results are as follows: (1) Aiming at the networked control system considering the modulation, the strategy of considering the bit error rate as the packet loss rate is put forward, and the blank of the problem of processing the modulation signal detection code in the field of the present networked control system is filled. In this paper, the angle of binary modulation and multi-system modulation is studied. For simpler binary modulation, the networked control system with the constant time delay and the error code is modeled as an asynchronous dynamic system. The stability analysis and the design of the state feedback controller parameters are carried out by means of the Lyapunov stability theory, and the bit error rate theorem of the networked control system with binary modulation is obtained; for the more practical and more complex multi-system modulation, The networked control system with the constant time delay and the error detection code is modeled as an asynchronous dynamic system. The stability analysis and the design of the dynamic output feedback controller are carried out by means of the Lyapunov theory, and the bit error rate theorem of the networked control system with the multi-system modulation is obtained. And the system selection theorem of the multi-system modulation is given. (2) For a networked control system with medium access constraints, the concept of control and scheduling coupling design is proposed for the first time, and based on the analysis of the control and scheduling coupling design method based on communication sequence and maximum error in the existing literature, A class least-square scheduling strategy based on the ideal dynamic of the system is proposed. After the ideal dynamic embedding of the system with no time delay is embedded in the scheduler, the state of the current state and the ideal dynamic state is compared with the state of the ideal dynamic state in each sampling period, the state with the large deviation preferentially obtains the authority of the access network, and the system analysis and the design conservatism are effectively reduced. On the basis of the proposed class least square scheduling algorithm, a networked control system with random short-time delay and medium access constraints is modeled as a switching system, and the coupling design of the class least-square scheduling and control is carried out by means of the theory of switching system and the Lyapunov stability theory. In addition, the proposed least-squares scheduling algorithm is further improved, and the ideal dynamics with time delay is replaced by the ideal dynamics without delay, and the real system can be more fully reflected due to the ideal dynamics with time delay, so the improved scheduling strategy has better performance. (3) Based on the model-based networked control system for saving the bandwidth, the method of the average residence time of the switching system is proposed to analyze and design the strategy, and the flexibility of the system analysis and design is improved. The model-based networked control system is modeled as a switching system, and the exponential stability analysis and the state feedback controller design are carried out on the model-based networked control system by means of the average residence time method of the Lyapunov stability theory and the switching control. In addition, the model-based networked control system architecture with an observer in the prior art is improved, and the observer is advanced to the controller by the access network, and the improvement not only saves the design cost but also avoids the problem of signal loss caused by network packet loss. After the model-based networked control system with the observer is modeled as a switching system, the exponential stability analysis of the model-based networked control system and the design of the controller and the observer are carried out using the Lyapunov stability theory and the mean residence time method of the switching control. (4) According to the model-based networked control system, the method of average residence time for pulse-switching control is proposed to analyze and design the strategy, and the conservative property of system analysis and design is reduced. The model-based networked control system is modeled as a pulse switching system by analyzing the phenomena of the system state transitions when the network is closed. The exponential stability analysis of the model-based networked control system and the design of the state feedback controller are carried out using the Lyapunov stability theory and the mean dwell time method of the pulse-switching control. In addition, on the basis of an improved model-based networked control system with an observer, a model-based networked control system with an observer is modeled as a pulse switching system, The exponential stability analysis of the model-based networked control system and the design of the controller and the observer are carried out using the Lyapunov stability theory and the mean residence time method of the pulse-switching control. (5) In view of the model-based networked control system, the model-based average residence time strategy for pulse switching control is proposed, and the conservative property of system analysis and design is further reduced. In the mode-dependent average dwell time strategy of the pulse-switching control, the average dwell time of the system depends not only on the switching parameters of the system but also on the pulse matrix parameters of the system. By analyzing and designing the model-based networked control system under the mode-dependent average residence time strategy of the pulse-switching control, a sufficient condition and a state feedback controller for the stability of the system index are obtained. In addition, the model-based networked control system with the observer is analyzed and designed by means of the average residence time strategy of the mode based on the pulse switching control, and the sufficient conditions for the stability of the system index and the design of the controller and the observer are given. In this paper, some examples are simulated by using the TrueTime toolbox of Matlab and the approximate network environment, and the validity of the proposed method is verified. Finally, the paper summarizes the work of the full text, and puts forward some suggestions on the work in need of further research.
【学位授予单位】：北京交通大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP273
，
本文编号：2490711
本文链接：https://www.wllwen.com/shoufeilunwen/xxkjbs/2490711.html
