The Use of Propagation Modellin in MUSIC and Madmum Likelihood for Loxdngle Tracking.
Edward F. Boss, J. Rody · 1990
The performance of the MUSIC algorithm as well as several other superresolution methods for spatial spectrum estimation degrades severely in a highly correlated signal environment as encountered in low-angle tracking involving multipath propagation. This deficiency can be overcome by replacing the direction-of-arrival (DOA) search vector ordinarily used with a special vector modelling the specular multipath. The model for specular multipath, named the refined model because of its use of additional a priori information, has been used previously in maximum likelihood (ML) estimation combined with multi-frequency illumination to give performance exceeding that of other high-resolution techniques for radar low-angle tracking in the presence of multipath. In this paper, we compare the performance of the MUSIC using a refined propagation model with ML using the same refined propagation model. The comparison is achieved by the means of Monte-Carlo simulations. The results indicate that both algorithms perform similarly for high signal-to-noise ratios but the ML approach is preferred because it is more efficient, requires less snapshots and gives better results for low signal-to-noise ratios.