Non-orthogonal Frequency Hopping Signal Underdetermined Blind Source Separation in Time-Frequency Domain

In this paper, a novel Matching Optimization Algorithm (MOA-algorithm) based on underdetermined blind source separation is proposed for non-orthogonal frequency hopping signal (that is, inner products are not always equal to zero in the same time-frequency point). Compared to traditional methods, the separation method is formulated as matching optimization. In our method, we accomplish the underdetermined blind source separation by computing the Short Time Fourier Transform (STFT) of each observation to get the signal timefrequency distribution, then we formulate the separation problem as matching optimization. In matching optimization, a new cost function is designed to improve the complete separation, and we make negative gradient direction as the steepest descent direction, to verify the proposed method on several simulations. The experimental results demonstrate the effectiveness of the proposed method.