An algorithm for blind adaptive multiuser detection based on hidden Markov models

The problem of constructing a blind adaptive multiuser detector (MUD) is considered for direct sequence code division multiple access signals transmitted through multipath channels. Due to the slow speed of convergence of channel estimation and existent unreason in state estimation in statistical algorithm based blind adaptive multiuser detection, hidden Markov model theory is proposed as the basis of a new algorithm which uses decorrelation to estimate the multichannel impulse response (MCIR). The algorithm extracts the correlation between the fore states and the back states and increases the speed of channel response estimation. Simulation results illustrate that it has an excellent effect in suppressing multipath interference and resolving the near-far problem, and, at the same time, the algorithm has global convergence and good stability. Under similar condition, the capability of the algorithm is better than that of MABW.