Group decision aided multiuser detection for overloaded multirate CDMA

In this paper, we present a multistage group-decision-aid MUD (MG-MUD) for the uplink multirate code division multiple access (CDMA) systems. To provide good tradeoff between the performance and computational complexity, the new MUD is based on the widely MMSE criterion and constructed by a bank of group detectors in each stage. First, the users are classified into different groups according to their transmission rates. Thereafter, these groups of users are detected in parallel by the set of the group detectors stage by stage. In addition, in each stage of the new MUD, the decision outputs of the group detectors are gleaned to refine the interference estimates to enhance the detection. To further reduce the computational overheads, the reduce-complexity MG-MUD is also presented, which divides the multiple symbols of the users in the higher rate groups into appropriate subgroups according to their transmission orders. Simulation results show that the proposed MUD can not only reduce computation load but provide a bit error rate (BER) performance close to the widely MMSE MUD, including in highly loaded and overloaded scenarios.