Blind multiuser detection based on modified stochastic gradient descent ICA for synchronous DS/CDMA signals in colored noise

In this paper, a blind adaptive multiuser detector based on stochastic gradient descent independent component analysis (ICA) method is proposed to suppress multiple access interference (MAI) over a flat Rayleigh fading channel with unknown colored noise. ICA based on blind source separation (BSS) technique suffers from inherent permutation ambiguity, which would cause the serious detection problem in direct sequence/code division multiple access (DS/CDMA) communications. In conjunction with the soft constraint of desired user's signature code, a modified stochastic gradient descent ICA-based blind multiuser detector can remove the inherent permutation ambiguity and achieve MAI suppression with small iterations in the presence of colored noise.