Method to separate electromagnetic interference sources based on underdetermined blind sources separation

Traditional testing methods of electromagnetic interferences can not observe individual airborne equipment when multiple devices are working. Furthermore, the existing blind sources separation algorithms can not solve the problem that the number of observed signals is less than the number of source signals. To overcome these shortcomings, a new underdetermined blind sources separation algorithm is proposed to separate electromagnetic interferences. The method is applied to harmonic signals with sparse characteristics. The algorithm constructs mathematical abstraction of electromagnetic interferences by underdetermined blind source separation mode. First, the single source area is found by calculating the ratio of observed sampling points. Then, the number of sources and mixture matrix are estimated using Hough-windowed method. Finally, the mixed signals are separated based on angle difference sorting method. Simulation results show that the effectiveness and accuracy of the proposed algorithm that the average correlation coefficient between separated signals and sources is 0.9936. Monte Carlo simulation results show the higher stability and noise immunity, and measured results demonstrate the feasibility of the algorithm.