Use of independent component analysis to separate noise sources in emissions radiated from LED lamps

摘要

Separation of noise sources from observations using an independent component analysis (ICA) allows the analysis of accumulated effects of noise radiated from equipment under tests, such as light-emitting diode (LED) lamps, which are sometimes simultaneously used in multiple numbers. The principal aim of this study is to experimentally perform blind source separation (BSS) by applying ICA to received signals containing multiple LED emissions in a time sequence. It is expected that this study will serve as preprocessing for investigations into the accumulated effects of LED emissions and thus will lead to an evaluation of interference in radio receptions. We report experimental results for LED emissions from two sources and two observations based on FastICA, which is one of the most commonly used algorithms in ICA source separation.