基于PPSO-MPC的多雷达协同反隐身指示搜索任务规划  

Mission Planning for Cued Search of Cooperative Anti-Stealth Detection Based on PPSO-MPC

作　　者：高晓光[1] 万开方[1] 李波[1] 李飞[1] 

机构地区：[1]西北工业大学电子信息学院,陕西西安710129

出　　处：《电子学报》2015年第9期1673-1681,共9页Acta Electronica Sinica

基　　金：国家自然科学基金(No.61305133);全国高校博士点基金(No.20116102110026);航天技术支撑基金(No.2013-HT-XGD);中央高校基本科研业务专项资金资助(No.3102015ZY092)

摘　　要：针对ESM/雷达协同反隐身探测中的指示搜索问题,引入模型预测控制(Model Predictive Control,MPC)理论,给出指示搜索任务规划的MPC框架,建立指示搜索的目标状态预测模型和在线滚动优化模型.针对模型求解,引入粒子群优化(Particle Swarm Optimization,PSO)算法,设计了高维矩阵粒子编码方式,引入尺度计算因子处理边界约束,引入概率模型处理离散变量,设计实现了一种"多主节点-单从节点"的(Multi-Master-Single-Slave,MM-SS)多种群并行计算策略.仿真结果表明,所建立的模型能够在不确定、多目标环境下实现对多雷达的高效协同控制,所提出的模型求解算法能够实现对滚动优化问题的快速、高效求解,即模型和算法的有效性得到了验证.To solve the cued search problem when ESMs and radars cooperate with each other in anti-stealth detection,a MPC-based（Model Predictive Control）mission planning frame for cued search is proposed,and the targets’states predictive model and on-line receding optimization model are established based on the MPC theory.Then,this paper puts forward an improved paral-lel PSO（Particle Swarm Optimization）algorithm to solve the problem.Concretely,a high-dimensional matrix mode is designed for particle coding,a scale-factor is imported for boundary restriction,a probabilistic model is proposed for processing discrete variable, and a new multi-swarm parallel strategy called MM-SS（Multi-Master-Single-Slave）is presented for promoting optimization effi-ciency.Experiments show that the established model realizes an efficient control of multi-radars in condition of uncertainty and mul-tiple targets,and that the proposed algorithm can solve the receding optimization problem efficiently.That is,the validity of the mod-el and algorithm is demonstrated.

关 键 词：反隐身 指示搜索 MPC 任务规划 滚动优化 PSO 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]