DOA estimation based on high efficient time-frequency analysis for multiple LFM signal sources

The performance of the DOA estimation approaches based on time-frequency analysis which have brought forward at present are not ideal for the multiple LFM signal sources situation. In this case, suppressing or eliminating cross interference terms is the key to improve direction estimation performance. This paper presents time-frequency DOA estimation based on SPW transform. The SPW combines the virtues of WVD and STFT, that is, it can suppress the cross terms quite effectively, and possesses high time-frequency aggregative-ness simultaneously. This method does not need oversampling and has the prominent advantages for reducing calculation cost. Simulation results demonstrate that in multi-sources situation, the approach based on SPW improves the DOA estimation precision much more greatly compared with the approaches based on PWVD and SP-WVD. For example, when the SNR is -20 dB, the root-mean-square errors of direction estimate of the approaches based on PWVD and SPWVD are 0.25° and 0.28°, respectively, while for the approach based on SPW, it is only 0.04°.