Authors : Mohcin Mekhfioui, Rachid Elgouri, Amal Satif, Aziz Benahmed, El-Mehdi Hamzaoui, Hadjoudja Abdelkader and Laamari Hlou
Abstract: The Blind Source Separation (BSS) consists in identifying a set of signals from an unknown source from a set of observed signals. It can be used for image processing, acoustics, telecommunications and biomedical applications. To perform the BSS different methods and properties are carried out. Independent Component Analysis (ICA) is frequently used. In our study, we are running a BSS algorithm on a portable environment the DSK TMS320C6713 card after the validation of this algorithm on MATLAB. The test signals are recorded by two microphones and sent to the DSK TMS320C6713 card via. a jack cable. By calculating the Signal to Interference Ratio (SIR) and the Performance Index (PI), we have proved that the algorithm SOBI (Second Order Blind Identification) of Belouchrani is the most efficient to analyze our signal database, so the separation is performed on the DSK TMS320C6713 using the SOBI algorithm to validate the its performance of the in real time.
Mohcin Mekhfioui, Rachid Elgouri, Amal Satif, Aziz Benahmed, El-Mehdi Hamzaoui, Hadjoudja Abdelkader and Laamari Hlou, 2020. A Comparative Study and Implementation of Blind Source Separation Algorithm using MATLAB and TMS320c6713 DSK. Journal of Engineering and Applied Sciences, 15: 1074-1081.