Distributed multiuser detection
Rick S. Blum, Jun Hu · 2003
Multiuser detection is extended to distributed scenarios. Each cooperating receiver makes preliminary decisions on the received bits of multiple users based on its own observations. Then the preliminary decisions are combined at the fusion center to produce the final decision. Optimum schemes, consisting of a fusion rule and the corresponding receiver decision rules, are found for various criteria. Three suboptimum schemes with reasonable complexity are also investigated. These are: the single-user matched filter, the decorrelating detector and the MMSE detector. For Gaussian noise and conditionally independent observation cases, the performance of the optimum and suboptimum distributed and centralized schemes are compared.