Joint active users' identification, timing estimation and multiuser detection in an asynchronous DS-CDMA system

This paper presents a novel Bayesian solution to the difficult problem of joint active users' identification, timing estimation and multiuser detection in an asynchronous DS-CDMA system. The observation vectors are considered as a compound Poisson process vitiated by the colored noise, the number of active users is subjected to Poisson distribution and active users' timing information followed the conditional distribution of arriving time in Poisson process. We show the conditional posterior probability density function by introducing subspace algorithm and Bayesian inference, then use RJMCMC algorithm and particle filter to parameters estimation and multiuser detection jointly. Simulation results support the effectiveness of methods.