Abstract: A new algorithm based on similarity measurement was proposed in order to address the issue of low estimation accuracy and high computational complexity in the existing algorithms for underdetermined mixing matrix blind estimation. The proposed algorithm can estimate the number of source signals and the mixing matrix automatically without any prior information or much iteration. Compared with the existing algorithms such as the K-means clustering and the Laplace's potential function method, the simulations turn out that the proposed has obvious advantages in the estimation accuracy of the number of source signals and the estimation precision of mixing matrix and the computational complexity.

Key words: underdetermined blind separation, blind estimation of the source number, estimation of mixing matrix, similarity measurement