无人机动态目标搜索的建模及求解  

Modeling of UAV's Dynamic Search and Its Realization Algorithm

作　　者：刘振[1] 胡云安[2] 史建国[3] 

机构地区：[1]海军航空工程学院接改装训练大队,山东烟台264001 [2]海军航空工程学院控制工程系,山东烟台264001 [3]海军航空工程学院七系,山东烟台264001

出　　处：《电光与控制》2013年第11期1-6,共6页Electronics Optics & Control

基　　金：国家自然科学基金(61174031;60674090)

摘　　要：对无人机(UAV)在动态战场环境中的协同搜索问题进行研究,在考虑目标存在概率、价值收益、UAV斥力代价及任务执行代价的情形下,建立了多UAV协同搜索的滚动优化模型。为了提高模型求解的效率,提出一种改进种群增量学习算法(PBIL)对该问题进行求解,采用混合编码的方法构造种群,同时采用了自适应的更新率,并利用自适应交叉和变异方式,将该算法应用于动态目标的搜索问题,仿真结果表明该方法能有效地搜索到战场目标,提高了搜索效率。The paper focuses on the problem of cooperative target search for multiple unmanned aerial vehicles （UAV） in dynamic battle environment.The model of cooperative search was established with consideration of target existence probabilitytarget valuerepulsive force and task execution cost.In order to enhance the solution efficiency of the modelan improved populationbased incremental learning （PBIL） was proposed.Hybrid encoding mechanism was adopted in the proposed algorithm for generating the populationand an adaptive updating rate was used.The crossover and mutation operator could also change adaptively according to the evolution condition.The improved PBIL was used to the dynamic target searchand the experimental results show that the algorithm can improve the search efficiency and can search more targets than before.

关 键 词：无人机 动态目标搜索 斥力 基于种群的增量学习 概率搜索图 

分 类 号：V279[航空宇航科学与技术—飞行器设计] TP18[自动化与计算机技术—控制理论与控制工程]