Research On Wideband Coherent Signal DOA Estimation Algorithm And FPGA Implementation

Wideband signals DOA estimations have become one of the main research topics of array signal processing.In engineering,with the increasing requirements for the implementation of algorithm confidentiality,DSP can not meet the needs for the implement of DOA estimation algorithm.The Field Programmable Gate Array can implement the algorithm from the bottom layer by using logic gate circuits,so more and more researchers are investing in FPGA implementation of algorithms.This thesis focuses on the DOA estimation algorithm for wideband coherent signals.Finally,the virtual array focusing algorithm is selected due to the estimation performance and the complexity of FPGA implementation,and the key steps for FPGA implementation are given.The main research work of this thesis is as follows:(1)Two classical algorithms of narrowband coherent signals,which are the spatial smoothing algorithm and Toeplitz matrix reconstruction algorithm,are researched in this thesis.The effectiveness of the DOA estimation of the two algorithms on coherent narrowband signals was verified by using MATLAB platform,and a comparative analysis was performed.At the same time,the implementation complexity of the two algorithms is evaluated from the perspective of hardware implementation.The results show that the overall operation complexity of the Toeplitz matrix reconstruction algorithm is lower than the spatial smoothing algorithm.This conclusion is used to solve the problem of DOA estimation of wideband coherent signals,and the Toeplitz matrix reconstruction algorithm and ISM algorithm are combined as a kind of algorithms for wideband decoherence problems.(2)For DOA estimation of wideband coherent signals,three types of algorithms are studied and analyzed.One is the combination of ISM algorithm and the algorithms of narrowband decoherence;the second is the classic wideband decoherence algorithm;the third is the Virtual array focusing algorithm which was newly proposed.Then,three types of algorithms were simulated under the same conditions by using MATLAB.According to the results of the simulation experiments,the DOA estimation result for wideband coherent signals from the RSS algorithm is the most accurate among the three algorithms.The estimating performance of the virtual array focusing algorithm is worse,and the estimating performance of the ISM algorithm combined with the Toeplitz matrix reconstruction algorithm is the worst.Finally,from the perspective of hardware implementation,the computation burden and the implementation complexity of three algorithms are analyzed.In combination with the estimating performance of the three algorithms previous simulated,the virtual array focusing algorithm is chosed to be the candidate algorithm for hardware implementation.(3)The algorithm theories involved in the FPGA implementation of the virtual array focusing algorithm were analyzed in detail.A 128-point FFT was implemented using the FFT IP core;a radix 2 in-situ FFT processor was designed to implement a 16-point FFT;the IDFT formula was used to implement the IDFT;a complex conjugate matrix was converted into a real symmetric matrix according to the related theory;the characteristic decomposition of a 16×16-dimensional real symmetric matrix was realized by using Jacobi algorithm and CORDIC algorithm;a circuit was designed based on the spectral peak function to realize spectral peak search.Each module was simulated using Modelsim,and the simulation results were compared with theoretical results to evaluate the correctness of the functions of each module.