Subspace tracking based blind adaptive multiuser detection for uplink CDMA

Abstract

In this paper, we first construct a blind multiuser detector based on block implementation, through batch eigenvalue decomposition (ED) of the sample autocorrelation matrix in asynchronous CDMA systems. It is computationally too expensive for adaptive applications, unsuitable to time-varying channels. We then propose a blind adaptive multiuser detection algorithm based on the PASTd algorithm for signal subspace eigencomponents estimation and a Kalman-type algorithm for channel estimation. The simulation results show that the proposed blind adaptive multiuser detection algorithm outperforms the RLS minimum-output-energy (MOE) blind adaptive multiuser detector and offers lower computational complexity.