稳定的有领航者的多移动agent群集运动控制  

Control of stable flocking motion of multiply-agent with a leader

作　　者：俞辉[1] 王永骥[1] 程磊[1] 

机构地区：[1]华中科技大学控制科学与工程系,湖北武汉430074

出　　处：《华中科技大学学报（自然科学版）》2005年第8期56-58,共3页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(69974017;60274020)

摘　　要：对具有二次积分动态的多移动agent跟随领航者(Leader)取得群集(Flocking)运动编队进行了研究.提出了一个分散控制方法对多移动agent进行分散控制,并通过理论证明得到以下主要结论:a.所有agent速度方向收敛到同一方向并与领航者保持一致;b.所有agent速度大小收敛并与领航者相同;c.互连的agent之间没有碰撞发生;d.所有agent的人工势场函数被最小化.用图论模型表示agent之间的相互作用及通信关系,对固定的网络拓扑,控制互连拓扑是固定的、时不变的,运用传统的李亚普诺夫理论进行了稳定性分析.最后,给出了一个计算机仿真例子对所得结论进行了验证.仿真结果表明,控制策略可以保证所有agent的速度大小和方向收敛到与Leader保持一致,同时避免碰撞,并保持一个紧凑的编队.The multiple mobile agents with double integrator dynamics followed by a leader to achieve flocKing motion formation were studied. A class of decentralized control laws for a group of mobile agents was proposed. From the theoretical proof, the following conclusions were reached： （i） global agent＇s velocity aligning their vectors with leader; （ii） convergence of their velocities to the leader＇s velocity; （iii） collisions among interconnected agents avoidance; （iv） the minimization of the agents＇ artificial potential function.The interaction and/or communication relationship among agents was modeled by graph theory. In fixed network topology of control interconnection being fixed and time invariant, the stability analysis was achieved by using classical Lyapunov theory. The simulation example was given, showing that the control policy ensured all agents eventually align with the leader and have the .same heading direction as the one of the leader agent while at the same time no collisions and group into a tight formation.

关 键 词：多智能体系统 协调控制 群集 图论 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]