Direction of arrival estimation of sources with intersecting signature in time–frequency domain using a combination of IF estimation and MUSIC algorithm

Time–frequency (TF) approaches are frequently employed for source localization at low signal to noise ratio. However, TF approaches fail to achieve the desired performance for sparsely sampled signals or signals corrupted by heavy noise in an under-determined scenario when sources are not TF separable. In this study, we propose a new TF method for direction of arrival (DOA) estimation of sources with closely spaced and overlapping TF signature. The proposed method uses a combination of a high-resolution time–frequency distribution and instantaneous frequency estimation method for extraction of sources with intersecting and closely spaced time–frequency signatures. Once sources are extracted, their DOAs are estimated using a well known multiple signal classification (MUSIC) algorithm. Experimental results demonstrate that the proposed source localization method achieves better performance as compared to the conventional time–frequency MUSIC algorithm.