Prediction-based PSO algorithm for MIMO radar antenna deployment in dynamic environment

作     者：Wang, Ziqin Zhang, Tianxian Kong, Lingjiang Cui, Guolong 

作者机构：Univ Elect Sci & Technol China Sch Informat & Commun Engn Chengdu Sichuan Peoples R China 

出 版 物：《JOURNAL OF ENGINEERING-JOE》 

年 卷 期：2019年第2019卷第20期

页      面：6646-6650页

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 08[工学] 

基　　金：Chang Jiang Scholars Program National Natural Science Foundation of China [61501083, 61771109] 

主　　题：MIMO radar particle swarm optimisation radar antennas optimisation autoregressive prediction model multiobjective particle swarm optimisation algorithm MIMO radar systems antenna deployment problem optimal algorithm multiple regions MIMO radar antenna deployment prediction-based PSO algorithm prediction strategy PSO method predicted solutions exact optimal solutions temporally optimal solutions previous information static problem time interval time period PSO optimisation current deployment schemes previous optimal information MOPSO-AR method dynamic environment 

摘      要：Under the circumstance of simultaneously scanning multiple regions in an environment which can change as time goes by, the authors study an optimal algorithm to solve antenna deployment problem for MIMO radar systems here. It is solved by multi-objective particle swarm optimisation algorithm (MOPSO) combining an autoregressive (AR) prediction model (MOPSO-AR). In a time period and dynamic environment, the MOPSO-AR method uses the previous optimal information to calculate the current deployment schemes before PSO optimisation starts up. It greatly reduces computational load and the error of the solutions. First, by discretising the time period into several time intervals, the problem in each time interval can be seen as a static problem. However, there may be relationship between these time intervals. Second, use the previous information and an AR model to predict temporally optimal solutions. Then, to get the exact optimal solutions, the predicted solutions and PSO method was applied to compute. Simulations show that the prediction strategy improves algorithm performance.