Abstract:

In the situation of noncooperative communication and military communication electronic warfare, multiple communication signals should be separated to extract some useful signals, which comes down to a problem of blind source separation of signal convolution. First the model of the received signals is built and transformed to the problem of joint block diagonalization. Then a new nonorthogonal joint block diagonalization algorithm is proposed. The new iterative algorithm is based on the steepest gradient descent method, and the optimization strategy and computation complexity are analyzed. Simulation results demonstrate the validity and reliability of the algorithm, which has a faster convergence speed than the existing methods.

Key words: communication signals;blind source separation of convolution;joint block diagonalization;steepest descent algorithm