Blind source separation of electromagnetic signals based on one-dimensional U-Net

Abstract

Digital electronic equipment emits electromagnetic signals under working conditions, resulting in information leakage and a serious threat to information security. To explore the extent of leakage of important information, blind source separation techniques are used to separate and detect mixed electromagnetic radiation signals. Deep learning techniques provide a feasible option for blind source separation detection of electromagnetic signals in noisy environments. In this paper, we use a one-dimensional u-net to blindly separate the electromagnetic signals leaked by the LCD display. Experiments show that the one-dimensional U-Net with five layers of ELU activation function has the best performance.