Blind Estimation of Underdetermined Mixing Matrix Based on Density Measurement

摘要

Under the circumstances that source signals are sufficiently sparse, an algorithm based on density measurement for blind estimation of the underdetermined mixing matrix is proposed in this paper. The proposed algorithm can estimate the number of source signals and the mixing matrix of the transmission channel simultaneously without any prior information. There are mainly three steps, including the preprocessing of observed samples, reservation of high-density samples, and estimation of the mixing matrix. Compared with the existing algorithms such as fuzzy clustering algorithm and probability density-based algorithm, the proposed algorithm does not require many iterations, which improves the efficiency. Simulation results show that the proposed algorithm has obvious advantages in the aspects of estimation accuracy of the mixing matrix as well as computational complexity and robustness.