摘要
研究了多智能体同时跟踪多个目标的群集算法,采用智能体的编号对目标数求余数的方法来确定各个智能体的跟踪目标。对于势场函数,根据智能体目标是否相同建立了两种势场函数,并在此基础上确定动态系统的控制输入,使相同目标的智能体在跟踪目标的过程中形成群集,而不同目标的智能体相互分离;用李亚普诺夫稳定性理论分析了算法的收敛性。仿真结果表明,所提算法可有效解决多目标跟踪问题。
A flocking algorithm of multi-target tracking for multi-agent is presented.By using the number of agents to get the remainders of the target numbers,the target for each agent is determined.Two kinds of potential functions are presented to get the inputs of the dynamic systems.Based on the potential functions agents with same target can make a flocking in the process of tracking,and the different ones will be separated from each other.Lyapunov stability theorem is applied to prove the algorithm convergene.Simulation results of two tracking targets of multi-agent show that the multi-agent tracking problem can be solved efficiently by the proposed method.
出处
《控制工程》
CSCD
2008年第4期420-422,共3页
Control Engineering of China
基金
国家自然科学基金资助项目(60525303
60704009)
燕山大学博士基金资助项目(B203)
关键词
多智能体
多目标
势场函数
群集
multi-agent
multi-target
potential function
flocking
