A novel modification of WSF for DOA estimation

This paper addresses the most basic and crucial problem in smart antenna, i.e., the estimation of DOA (Direction-of-Arrival) finding. The performance of smart antenna system greatly depends on the resolution of DOA. MUSIC (MUiltiple SIgnal Classification) and ESPRIT (Estimation of Signal Paramter via Rotational Invariance Technique) are the most classic two algorithms for DOA finding in real systems. However, these two algorithms cannot handle coherent signals directly which happens for example in multipath propagation and the performance will be greatly deteriorated if the pre-processing technique such as spatial smoothing is used. Therefore, the system employing these two algorithms usually works in the condition of Line-of-Sight (LOS), e.g., in suburb circumstance. WSF (Weighted Subspace Fitting) algorithm is a more superior technique which has much higher resolution and can handle coherent signals without any pre-processing. However, conventional WSF needs to know the independent number of signals, otherwise its performance will be deteriorated. In this paper, we propose a modified WSF algorithm for DOA. The proposed modified WSF can detect the independent number of signals automatically and show much higher resolution compared to conventional WSF and MUSIC.