Source Number Estimation Based on Improved Singular Value Decomposition at Low SNR

Source number estimation is one of the key problems in blind signal separation. In order to solve the problem of weak anti-noise ability of the current method, this paper improves the singular value decomposition algorithm, converts the ratio problem into the included Angle problem, and effectively detects the turning point between the singular value and the noise platform. Theoretical analysis and simulation results show that the proposed algorithm is more robust when SNR is low.