Research On WSF Algorithm For Low Complexity Based On Particle Swarm Optimization

DOA estimation(DOA)is widely used in communication,radar,sonar,positioning,navigation fields.Weighted Subspace Fitting(WSF)algorithm have better accuracy in DOA,and is independent with multipath and coherent signal sources.However this algorithm is of very difficult in realization and in the amount of computation,which can't guarantee real-time communication.In a world,the simplification of this algorithm is becoming a key sport in DOA research.In this paper,we have studied the former researches on DOA estimation,and designed the Two Segment Weighted Subspace Fitting(TSWSF),and improved it based on it's disadvantages,proposed the Limited Solution Space Weighted Subspace Fitting(LSWSF)algorithm to lower the computational complexity.The former optimization ways are focused on the operator of algorithms or the use of faster intelligent optimization algorithms,although these algorithms introduced intelligent optimization algorithms,these methods still ignored the characteristics of mathematical model,which makes it possible to optimize further.To solve these problems,This paper preprocess the solution space firstly,then introduced correction parameters based PSO algorithm to solve the problem.The compression of solution space reduced the computational complexity.In order to further reduce computational complexity of this algorithm,continuous limit solution space solving strategy is added based on the inadequate of complexity increasement.The simulation and comparative analyze with former WSF optimization algorithms are based on MATLAB,which indicates that the proposed algorithm have more advantages both on computational complexity and on iteration times under the same DOA estimation accuracy.What's more,The good performance of LSWSF(compared with TSWSF)algorithm indicates the effectiveness of limit solution space strategy.