Research On High Resolution DOA Estimation Algorithms For Wideband Signal

The number of radar, sonar, and wireless communication applications requiring thelocation of a wideband signal source is ever increasing. As a result, many algorithmshave been proposed to improve DOA (direction-of-arrival, DOA) estimationperformance and computational efficiency. Based on uniform linear array model, thispaper presents several high resolution algorithms for wideband signal. The main work issummarized as follows:First, two classical signal subspace methods, MUSIC and ESPRIT, are introducedaccording to the mathematical model for narrow band array signal. The simulationsshow that the performance of these algorithms changes with SNR and the number ofsnapshots.Moreover, as an essential difference between narrow band and wideband arraysignal processing, the wideband signal model in frequency domain has been worked out.Simulations based on this model show that the performance of MUSIC changes withdifferent bandwidth. Additionally, three types of wideband signal DOA estimationalgorithms have been introduced which include ISSM, CSSM and wideband directprocessing. Also, beam space method has been applied to CSSM to obtain a lower SNRresolution threshold. According to the simulation results, the theoretical conclusionshave been verified.Finally, a new type of DOA estimation algorithm is introduced, which is called testof orthogonality. Two algorithms based on this approach are test of orthogonality ofprojected subspaces (TOPS) and test of orthogonality of frequency subspaces (TOFS).The simulations show that this technique performs better than others in mid SNRranges.