基于MAS的多UAV协同任务分配设计与仿真

Design and Simulation of Multi-UAV Coordinated Task Allocation Based on MAS

对多无人机UAV(Uninhabited Aerial Vehicle)协同任务问题,提出了基于多智能体系统MAS(Multi-Agent System)的系统框架;对多UAV协同目标分配问题建立了数学模型,提出基于Agent的分布协同拍卖的动态任务分配算法,通过多Agent拍卖实现目标分配;建立了多UAV仿真系统,基于HLA/RTI和知识共享与操作语言KQML(Knowledge Queryand Manipulate Language)实现多UAV系统中各Agent的交互;仿真系统能方便有效的仿真多机协同问题,对目标分配问题的仿真结果表明,目标分配算法具有良好的优化效果和时间特性,能够满足一般战场的实时性需求。

To study problems of multi-Uninhabited Aerial Vehicle （UAV） coordinated task problem, a multi-UAV coordinated task structure based on Multi-agent System （MAS） was proposed. The mathematic model for multi-UAV cooperative target assignment problem was build, and a dynamic target assignment algorithm based on multi-agent decentralized cooperative auction was presented, which assigned targets to multi-UAVs through a process of multi-agent auction. To verify the multi-UAV system, a simulation system was developed. The High Level Architecture/Run Time Infrastructure （HIA/RT1） and the Knowledge Query and Manipulate Language （KQML） were used to accomplish interactions between agents. The simulation system facilitated simulation problems of multi-UAV coordination. The simulation results of the target assignment problem demonstrate that the algorithm has good optimal effect and time performance, The algorithm could meet the common battle field requirement.