DOA Estimation of Wideband Signal Based on Improved MUSIC Algorithm

Abstract

The existing Direction of Arrival（DOA）estimation algorithms estimate the wideband signal azimuth in the passive detection system，there are some problems such as large deviation of DOA estimation results and high computational complexity，which are difficult to meet the requirements of signal real-time processing. To improve the spatial resolution of DOA estimation of multisource signals，this paper proposes an improved multiple signal classification algorithm based on the S-transform that does not require the estimation of the number of signal sources. According to the frequency-domain characteristics of wideband signals，the S-transform is used to process the received signal of the array to establishthe multi-resolution time spectrum matrix.At the same time，an array signal data model in the time-frequency domain is constructed.Based on the characteristics of the joint diagonalization structure of the signal power spectrum matrix，a subspace spectrum estimation equation based on the S-transform is formulated. On this basis， the DOA is estimated through a spectral search to realize the sound source location of multisource wideband signals.The simulation results show that the estimation success rate of the algorithm is always greater than 90% when the signal-to-noise ratio range is approximately-15~10 dB. Compared to the TCT and CS_TCT，CWT_MUSIC algorithms，the proposed algorithm has better estimation performance and does not need to estimate the number of signal sources.