Blind source separation techniques for interference sources detection in electromagnetic environments of ethernet networks

• Published in: AIP conference proceedings Vol. 2591; no. 1
• Main Authors: Rasheed, Ameer Thabit, Rashid, Hasan Th, Al-Bakri, Ali A., Abdullah, Ahmed Kareem
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Melville American Institute of Physics 29.03.2023
• Subjects: Algorithms; Audio data; Crosstalk; Ethernet; Independent component analysis; Interference; Signal to noise ratio; Stone
• ISSN: 0094-243X; 1935-0465; 1551-7616; 1551-7616
• DOI: 10.1063/5.0128512

The transmission process of sending data across carrier media whether it's text, audio, and graphics or any other source of data is influenced by the environment in which it's sent, and the transmitted signals are distorted as a result. Blind source separation BSS is a popular technique for recovering signals from separate sources that are isolated from the environment. In this paper, three poplars algorithms based on BSS techniques are used to identify sources of interference in an electromagnetic environment. Audio, text, and image data are utilized across networks with interference and noise, where the original data and another two types of Electromagnetic and Crosstalk noises were created; then merged and delivered over a network through Ethernet cable. The stone algorithm, Joint Approximate Diagonalization of Eigenmatrices JADE Algorithm, and the Enhanced Fast Independent Component Analysis EFICA algorithm are employed to exposing the information and the data passing through the cable for detecting the harmful impacts of interference sources over the electromagnetic environment. As a result, subjective and objective evaluation is done based on the popular criteria of signal to noise ratio SNR where the stone algorithm has best value of SNR over the others with 33 while JADE algorithm has 25, and EFICA algorithm has 4.