Embedded fastICA algorithm applied to the sensor noise extraction problem of foundation fieldbus network

Abstract

This paper presents the description and the operation of a system composed of an intelligent algorithm, that separates information and noise coming from different sources, implemented with embedded technology in a DSP (digital signal processor), that interacts with field bus devices connected through a foundation field bus network. The technique used in this blind source separation (BSS) process was the independent component analysis (ICA), that explores the possibility of separating mixed signals based on the fact that they are statistically independent. The algorithm and its implementation are presented, as well as the test results.