The Structure Optimization Of Subarray In Hybrid MIMO Phased-array Radar

Compared with phased-array radar，the Multiple-Input Multiple-Output(MIMO)radar adopting orthogonal waveform enjoys waveform diversity gain by sacrificingthe coherent processing gain. To resolve this problem, some scholars abroad haveproposed a new radar system called Hybrid MIMO Phased-Array Radar(HMPAR),which combines the advantages of the MIMO radar with the Phased-Array radar.In an HMPAR system, the transmit array is partitioned into a number ofsubarrays which transmit mutually orthogonal waveform. By designing a weightvector for each subarray, we can form a beam towards a certain direction in space,and each subarray can constitute the MIMO radar model. The coherent processingwithin each subarray and the orthogonality of waveforms between the subarrayscontribute to achieving the coherent processing gain and the waveform diversitygain.In this thesis, we propose a method to use one plane array for both transmittingand receiving, and the signals at different sub-array are orthogonal while withineach sub-array are coherent, then it adopts the sub-array processing in both thetransmitter and the receiver to achieve the intergration of phased-array radar andMIMO radar.The primary content in this paper involves: deducing the formulations of thetransmit/receive and the waveform diversity beampatterns by formulating the signalmodel for the hybrid MIMO phased-array radar; decoding scheme and subarrayconfiguration constraints; objective function and fitness function of subarrayoptimizing; optimal structure designing of subarray by using of multi-objectivegenetic algorithms.Since the HMPAR system model is based on phased-array radar and MIMOradar, we introduce the traditional phased-array radar and MIMO radar in thefollowing order: transmition-getting echos-receiving and processing. After that, itshows the derivation of signal processing procedure of2d HMPAR system model atsubarray level.In the decoding scheme, it adopts the chromosome coding, which encode thestructure of plane array into vector in the form that can be processed bymultiobjective evolutionary algorithm. Moreover, to concentrate the array structurerelatively, it assumes that the array follows the constrains that are not overlap andfulfilled.We set multiple objective functions, which respectively shows the performanceof waveform diversity, coherent processing gain and the transmitting/receiving direction pattern.Two kinds of multi-objective genetic algorithms have been useded in theoptimal structure designing of subarray, that including Pareto rank and weightcoefficient method. Then we compare the performance of each meach method onoptimation of objective function, and then the optimal array structure, that thesubarrays are divided by regular and relatively concentrated, can be achieved at last.