Performance Evaluation of Blind Source Separation Schemes for Separating Sensor Signals in a Distributed Network

In this paper we consider the problem of sensing the mixture of sources in a distributed network of sensors. We assume that sensors have low computation power. They quantize the mixtures and send them to the server for separation. We investigate the system performance using two low complexity blind source separation methods, FastICA and JADE. FastICA works based on minimization of negentropy of mixture signals, and JADE works based on diagonalizing covariance matrix of two the mixture signals. For simulation we mixed two memoryless Laplacian sources. We encode the mixtures using entropy coded deadzone quantizer at different bit-rates, and separate their encoded mixture signals. We compare the mean square error of reconstructed sources after separation. Results show that JADE has better performance especially for mixtures encoded at low bit-rates.