分数阶多智能体系统一致性控制研究
发布时间：2018-10-15 19:53
【摘要】：整数阶系统在描述带有记忆性材料和黏滞性材料会出现不稳定和不能反映其本身性能的局限,科学家们发现分数阶系统描述这类材料时稳定性和本身性能得到了很好的反映,因此近几年科学家们开始研究分数阶系统,并在分数阶机器人和电路有着广泛的应用.一致性问题是多智能系统研究的一个热点也是一个最基本的问题,主要通过多智能体之间的信息交流,根据它们之间的信息交流设计控制器,使得系统的状态到达一致.本硕士论文主要研究分数阶多智能体系统一致性问题,论文的安排如下:第一章介绍了分数阶系统的背景发展以及应用前景,并且也对分数阶微积分进行了简单的介绍.以及对本文所需的预备知识进行了介绍.第二章研究了有向图下分数阶多智能体系统基于观测器的一致性.通过假设状态信息不可测量而输出信息可以测量,构造三种类型的基于观测器的一致性协议.通过图论,矩阵论和李亚普诺夫方法等数学知识,得到了分数阶多智能系统到达一致的充分条件.第三章研究了带有领导的分数阶多智能系统追踪问题.每个分数阶动力系统都是同质的和包含未知的非线性项.在拓扑图是无向的,非线性项能通过神经网络参数化的假设条件下,运用自适应学习的方法处理非线性动力系统,基于此构造协作跟踪协议.通过解出黎卡迪方程得出反馈增益矩阵.同时构建一个完全分布式的自适应协议,自适应率用来调节耦合权重.设计相应的控制率,并且证明了所有闭环信号是一致有界的.第四章研究了在有向拓扑下的完全分布式的不确定非线性系统的追踪问题.基于状态信息的分布式协议下,利用神经网络去学习不确定非线性项,最终跟踪到了领导.同时也设计了基于观测器的跟踪协议.最终在此协议下跟踪到了领导.得出了所有闭环信号都是一致有界的结论,并且追踪误差和基于观测器的追踪误差收敛到一个很小的邻域.第五章论文小结与展望。
[Abstract]:Integer order systems describe the instability of materials with memory and viscosity and the limitations that do not reflect their own properties. Scientists have found that the stability and properties of fractional order systems in describing such materials are well reflected. In recent years, scientists have begun to study fractional order systems, and have been widely used in fractional robots and circuits. The consistency problem is a hot topic and the most basic problem in the research of multi-intelligent system. The controller is designed according to the information exchange between the multi-agents to make the state of the system consistent. In this thesis, the consistency of fractional multi-agent system is studied. The arrangement of the thesis is as follows: chapter 1 introduces the background development and application prospect of fractional order system, and also introduces fractional calculus briefly. And the preparatory knowledge needed in this paper is introduced. In chapter 2, the observer-based consistency of fractional-order multi-agent systems under directed graph is studied. Based on the assumption that state information is unmeasurable and output information can be measured, three types of observer based consistency protocols are constructed. By means of mathematical knowledge such as graph theory matrix theory and Lyapunov method sufficient conditions are obtained for fractional multi-intelligent systems to reach consistency. In chapter 3, the problem of fractional order multi-intelligent system tracking with leadership is studied. Every fractional dynamical system is homogeneous and contains unknown nonlinear terms. Under the assumption that the topological graph is undirected and the nonlinear term can be parameterized by neural network, the adaptive learning method is used to deal with the nonlinear dynamic system, based on which a cooperative tracking protocol is constructed. The feedback gain matrix is obtained by solving the Riccardy equation. At the same time, a fully distributed adaptive protocol is constructed, and the adaptive rate is used to adjust the coupling weight. The corresponding control rate is designed and it is proved that all closed loop signals are uniformly bounded. In chapter 4, the tracing problem of fully distributed uncertain nonlinear systems under directed topology is studied. Under the distributed protocol based on state information, neural network is used to learn the uncertain nonlinear term, and finally the leader is tracked. An observer based tracking protocol is also designed. Finally, the leader was tracked under this agreement. It is concluded that all closed-loop signals are uniformly bounded, and the tracking error based on observer converge to a small neighborhood. The fifth chapter is the summary and prospect of the thesis.
【学位授予单位】：温州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18;TP13
[Abstract]:Integer order systems describe the instability of materials with memory and viscosity and the limitations that do not reflect their own properties. Scientists have found that the stability and properties of fractional order systems in describing such materials are well reflected. In recent years, scientists have begun to study fractional order systems, and have been widely used in fractional robots and circuits. The consistency problem is a hot topic and the most basic problem in the research of multi-intelligent system. The controller is designed according to the information exchange between the multi-agents to make the state of the system consistent. In this thesis, the consistency of fractional multi-agent system is studied. The arrangement of the thesis is as follows: chapter 1 introduces the background development and application prospect of fractional order system, and also introduces fractional calculus briefly. And the preparatory knowledge needed in this paper is introduced. In chapter 2, the observer-based consistency of fractional-order multi-agent systems under directed graph is studied. Based on the assumption that state information is unmeasurable and output information can be measured, three types of observer based consistency protocols are constructed. By means of mathematical knowledge such as graph theory matrix theory and Lyapunov method sufficient conditions are obtained for fractional multi-intelligent systems to reach consistency. In chapter 3, the problem of fractional order multi-intelligent system tracking with leadership is studied. Every fractional dynamical system is homogeneous and contains unknown nonlinear terms. Under the assumption that the topological graph is undirected and the nonlinear term can be parameterized by neural network, the adaptive learning method is used to deal with the nonlinear dynamic system, based on which a cooperative tracking protocol is constructed. The feedback gain matrix is obtained by solving the Riccardy equation. At the same time, a fully distributed adaptive protocol is constructed, and the adaptive rate is used to adjust the coupling weight. The corresponding control rate is designed and it is proved that all closed loop signals are uniformly bounded. In chapter 4, the tracing problem of fully distributed uncertain nonlinear systems under directed topology is studied. Under the distributed protocol based on state information, neural network is used to learn the uncertain nonlinear term, and finally the leader is tracked. An observer based tracking protocol is also designed. Finally, the leader was tracked under this agreement. It is concluded that all closed-loop signals are uniformly bounded, and the tracking error based on observer converge to a small neighborhood. The fifth chapter is the summary and prospect of the thesis.
【学位授予单位】：温州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18;TP13
【相似文献】
相关期刊论文 前10条
1 王振滨,曹广益,朱新坚;一类分数阶系统的辨识算法(英文)[J];Journal of Southeast University(English Edition);2004年03期
2 赵春娜;薛定宇;;一种分数阶线性系统求解方法[J];东北大学学报(自然科学版);2007年01期
3 赵春娜;张祥德;孙艳蕊;;成比例分数阶系统的仿真研究[J];系统仿真学报;2008年15期
4 周亚非;王中华;;分数阶混沌激光器系统的同步[J];半导体光电;2008年05期
5 朱呈祥;邹云;;分数阶控制研究综述[J];控制与决策;2009年02期
6 左建政;王光义;;一种新的分数阶混沌系统研究[J];现代电子技术;2009年10期
7 汪纪锋;肖河;;分数阶全维状态观测器设计[J];重庆邮电大学学报(自然科学版);2009年06期
8 孙克辉;杨静利;丘水生;;分数阶混沌系统的仿真方法研究[J];系统仿真学报;2011年11期
9 李安平;刘国荣;沈细群;;不同阶分数阶混沌系统的同步与参数辨识[J];计算机工程与应用;2013年04期
10 严t，
本文编号：2273665
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2273665.html
