Multiuser Detection with Branch-and-Bound-Based Algorithm Using User Grouping for Synchronous CDMA

Inhalt:
In this paper, we propose a novel method for multiuser detection (MUD) based on the branch-and-bound (BBD) method using user grouping. The proposed method partitions K users into M disjoint groups and applies the BBD method to each group. Due to this user grouping, the number of users for the BBD method applied to a group is less than K. Therefore, the worst-case computational cost of the proposed method can be reduced dramatically compared with that of the BBD method. For an improvement of the probability of group detection error (GDE), the user signals of other groups are removed in parallel by using a solution of the decorrelating decision-feedback detector (DDFD) that is applied to all users before the user grouping. Moreover, P candidate solutions are derived in each BBD method. Finally, the best solution is determined from all combinations of the P candidate solutions. It is clarified the worst-case computational cost of the proposed method is much less than that of the BBD method, which is within the possible range to be implemented in a practical system. The simulation results show that the average computational cost of the proposed method is same order as that of the BBD method and the DDFD. They also denote that the probability of GDE of the proposed method is larger than that of the BBD method, but smaller than that of the DDFD.