具有约束条件的多智能体系统一致性与编队控制研究

发布时间：2017-12-31 05:34

本文关键词：具有约束条件的多智能体系统一致性与编队控制研究　出处：《燕山大学》2016年硕士论文　论文类型：学位论文

【摘要】：多智能体系统凭借其在计算机网络、人工智能、通信工程、编队控制、搜救等方面的广泛应用,已经引起了许多学者的研究兴趣。一致性控制是多智能体系统研究的基本问题之一,是多智能体系统其他问题研究的前提。然而实际环境中存在的许多不可测因素往往会直接影响多智能体系统是否可以达成期望的一致性。因此如何使多智能体系统在有约束的条件下形成一致性和编队的研究就变得非常有意义。首先,在有输入饱和的约束下,研究了线性多智能体系统基于分布式事件驱动一致性控制协议。提出的分布式的事件触发条件,有效的降低了智能体的控制更新次数,从而降低了能量消耗。进一步结合提出的事件触发条件设计了输入饱和约束下的多智能体系统的一致性控制协议,最终用S-procedure方法给出了该事件触发控制协议的收敛性条件。其次,考虑实际环境中智能体之间不可避免的通讯时延,进一步研究了在输入时延的约束下,基于事件触发的多智能体系统饱和一致性控制问题。在前述研究基础上将S-procedure方法拓展到了时滞系统当中,并进一步给出了该控制协议的收敛性条件。最后,在一致性理论研究的基础上,结合移动传感器网络覆盖控制问题,设计了移动传感器运动过程中的Leader-follower编队覆盖控制协议。在领导者信息不可测的约束条件下,对虚拟领导者设计了引力势函数,使其能够顺利到达目标覆盖位置。对跟随者设计了速度和加速度观测器,设计了编队控制协议。在该控制协议下,移动传感器能够在到达覆盖区域之前迅速的形成期望的覆盖编队,从而提高了传感器的覆盖效果和灵活性。
[Abstract]:Multi-agent system is widely used in computer network, artificial intelligence, communication engineering, formation control, search and rescue and so on. Conformance control is one of the basic problems in the research of multi-agent system. However, many untestable factors in the actual environment will directly affect whether the multi-agent system can achieve the desired consistency. Therefore, how to make the multi-agent system achieve the desired consistency? It is very meaningful to study the formation of consistency and formation under constrained conditions. First of all. Under the constraint of input saturation, the distributed event-driven consistency control protocol for linear multi-agent system is studied. The distributed event trigger condition is proposed, which can effectively reduce the control update times of the agent. The proposed event trigger condition is used to design the consistency control protocol for multi-agent systems with input saturation constraints. Finally, the convergence condition of the event-triggered control protocol is given by using S-procedure method. Secondly, the inevitable communication delay between agents in the real environment is considered. The input delay constraint is further studied. Based on the above research, the S-procedure method is extended to time-delay systems. Furthermore, the convergence conditions of the control protocol are given. Finally, based on the research of consistency theory, the coverage control problem of mobile sensor networks is combined. The Leader-follower formation overlay control protocol is designed in the process of moving sensors. The gravitational potential function is designed for the virtual leader under the constraint of untestable leader information. The speed and acceleration observer is designed for the follower and the formation control protocol is designed under the control protocol. The mobile sensor can quickly form the desired overlay formation before reaching the coverage area, thus improving the coverage effect and flexibility of the sensor.
【学位授予单位】：燕山大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP18;TP212

【相似文献】