Abstract:

The model inherent connections between underdetermined blind source separation and compressed sensing is analyzed, and the mathematical model of underdetermined blind signal reconstruction problems based on the compressed sensing model is built. The mixing matrix is estimated using the structure of the time domain or transform domain of the source signals. Based on the estimated mixing matrix, the greedy  orthogonal matching pursuit method is used to realize the reconstruction of the underdetermined sparse source signals. The proposed method can be modified to satisfy different applications. Numerical experiments including the comparison with a recent underdetermined blind source separation approach are provided to show the effectiveness of the proposed method.