通信约束下的分布式系统一致性控制方法研究

发布时间：2018-06-09 00:50

本文选题：分布式系统 + 通信约束　；参考：《上海大学》2016年博士论文

【摘要】：随着网络化、信息化与自动化技术的蓬勃发展和交汇融合,许多分布式系统扩展出更为强大和灵活的功能。为保证系统稳定运行,组成单元间需要通过网络通信交互信息协同合作。然而,数据在经输送能力有限的网络环节交互过程中,可能会出现通信网络数据传输量大、链路拥塞、多源噪声干扰等现象,从而影响控制效果甚至威胁系统稳定运行。为实现分布式系统有效控制,本文首先针对网络通信环节,研究可应用于路由节点的拥塞控制方法和传感节点的事件触发策略,保证通信链路通畅、降低网络数据传输量,以提高通信网络数据传输质量。考虑到通信网络中的不确定性因素难以避免,本文进一步考虑数据传输时变时滞、数据包随机丢失、通信拓扑动态切换等因素,研究具有容忍通信约束能力的分布式一致性控制策略,以期在保证系统稳定运行条件下提高控制效果。主要研究内容及创新工作如下：(1)现有的网络路由拥塞控制策略往往存在计算量大、鲁棒性差等缺点,在充分考虑状态不可测、链路受干扰、节点计算负荷重等因素的情况下,首先提…具有鲁棒性的观测器设计方法,以估计系统运行状态。然后,设计具有低计算负荷的路由节点缓冲队列控制策略,从维持节点队列稳定的角度保障通信链路数据传输通畅。(2)分布式系统中通常通过分散布置的传感节点获取信息,但其中冗余数据会加剧通信和控制计算压力。为降低数据发送量以缓解通信链路数据传输量,首先采用具有二次型的事件触发策略以剔除冗余数据,然后面向由同构节点组成的分布式系统一致性目标,考虑通信时变延时和随机拓扑切换因素,提出基于事件触发的分布式协同控制策略；进而以异构节点组成的分布式系统一致性为目标,创新提出指数型事件触发器的分布式补偿控制策略,可通过设定触发器参数以避免控制器短时频繁动作,最终保证系统轨迹收敛于要求稳定域。(3)分布式系统状态在通过仪表量测和网络化信息传输的过程中,容易受到来自量测和网络方面的多源、复杂干扰影响,同时可能会出现传输时变延时和拓扑随机切换现象,导致关键数据出现错位甚至错误,从而降低系统控制性能。为实现上述多通信不确定性影响下分布式系统一致性目标,首先设计点对邻域抗噪H指标,以提高系统抗多源复合噪声能力；进而设计具有容忍通信不确定性因素能力的分布式协同控制策略,所得的稳定性条件不需要额外的补充条件,从而降低控制器求解约束条件,同时保证系统均方一致性。(4)实际中的许多大规模分布式系统组成节点数量多、节点问存在本质的耦合关联,而现有一致性策略较少考虑控制器负荷,缺乏考虑节点趋同目标的差异性和利用系统本身的自治能力。因此,首先为筛选系统关键控制节点,提出控制关键节点选择依据；同时为降低控制器负荷,提出基于事件触发的控制命令更新策略；进而考虑链路噪声和通信拓扑动态切换因素,提出容忍通信不确定性因素的一致性牵制控制策略,通过控制少量节点、发布少量控制命令,充分利用节点自治能力保证系统群组一致性。
[Abstract]:With the networking, the rapid development and convergence of information and automation technology, many distributed systems have extended more powerful and flexible functions. In order to ensure the stable operation of the system, the intercommunication information of network communication is required to cooperate with each other. However, data can be used in the interaction process of a network link with limited transport capacity. In order to realize the effective control of the distributed system, this paper first studies the congestion control method which can be applied to the routing nodes and the event trigger strategy of the sensor nodes in order to achieve the effective control of the distributed system. In order to ensure the smooth communication link, reduce the network data transmission and improve the quality of communication network data transmission, considering the uncertainty in the communication network, this paper further considers the time delay of data transmission, the random loss of data packets, the dynamic switching of communication topology and so on. The main research content and innovation work are as follows: (1) existing network routing congestion control strategies often have shortcomings such as large computation, poor robustness, and so on, considering factors such as unmeasurable state, link interference and load weight of nodes. Under the circumstances, first of all... A robust observer design method is used to estimate the running state of the system. Then, a routing node buffer queue control strategy with low computing load is designed to ensure the smooth data transmission of the communication link from the point of maintaining the stability of the node queue. (2) the distributed system usually passes the distributed sensing nodes to obtain information, but among them Redundant data can aggravate communication and control computing pressure. In order to reduce the amount of data transmission to alleviate the communication link data transmission, first of all, the two type event trigger strategy is used to eliminate redundant data, and then the distributed system conformance target is composed of isomorphic nodes, and the time delay and random topology switching factors are considered. The distributed cooperative control strategy based on event triggering is proposed. Then the distributed compensation control strategy of exponential event trigger is put forward with the goal of distributed system consistency composed of heterogeneous nodes. It can avoid the short and frequent motion of the controller by setting the trigger parameters, and ultimately ensure that the system trajectory converges to the required stability. (3) in the process of measurement and network information transmission, the state of distributed system is easily affected by multi source and complex interference from measurement and network. At the same time, time delay and topology random switching may occur, resulting in the misplacement and even error of the key data, thus reducing the performance of the system. In order to achieve the conformance target of the distributed system under the influence of the above multi communication uncertainty, the neighborhood anti noise H index is designed to improve the system's ability to resist multi source compound noise, and then the distributed cooperative control strategy with the ability to tolerate the uncertainty of communication is designed. In order to reduce the constraints of the controller to solve the constraints, and ensure the uniformity of the system. (4) many large-scale distributed systems in reality make up the number of nodes, and the nodes ask the essential coupling Association, while the existing consistency strategies less consider the controller load, lack the difference in considering the node convergence target and use the system itself. Therefore, in order to screen the key control nodes of the system, the selection basis of control key nodes is proposed. At the same time, a control command updating strategy based on event triggering is proposed to reduce the load of the controller, and then the link noise and the dynamic switching factors of the communication topology are considered, and the consistency control of the uncertain factors of tolerance communication is proposed. By controlling a small number of nodes, a small number of control commands are issued to make full use of node autonomy to ensure system group consistency.
【学位授予单位】：上海大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP273

【相似文献】