Blind source separation algorithm for colocated MIMO radar signals

In electronic countermeasure, reconnaissance and signal sorting of colocated MIMO radar is a difficult point. The blind source separation algorithm based on noncircular complex fast ICA(NC-Fast ICA) is utilized to sort MIMO radar signals. In order to satisfy the application conditions of the BSS algorithm, a reconnaissance method for collecting radar signals in multiple positions with different angles of view is proposed to obtain and a separable full-rank mixing matrix. Several typical orthogonal waveforms such as phase-coded signals and discrete frequency coded waveforms are used for separation. Simulation results show that a separation performance of -14dB can be achieved under the condition of evenly angle spaced sampling and a signal to noise ratio of 10dB regardless of the amplitude fluctuations. The correlation comparison and time-frequency analysis of separated waveforms also demonstrate the validity of proposed method.