NMF-based Localization of Spatially Sparse Overlapping Coherent Distributed Sources

In this paper, we consider the problem of locating spatially distributed sources in a complex scenario where two or more coherent sources overlap. Based on the convex optimization criterion of sound source location based on sparsity constraint, the non-negative matrix factorization (NMF) technique is introduced to realize the non-negative estimation of the spatial distribution function of distributed sources and improve the estimation accuracy. In this paper, simulation experiments are carried out under different configurations, and the results show that the method is effective.