Reduced Complexity Distributed Base Station Processing in the Uplink of Cellular Networks

We propose a reduced-complexity belief propagation (RCBP) receiver for joint detection in the uplink of cellular multi-access networks with base station cooperation. Using local message passing algorithms based on belief propagation (BP) has been shown to be very promising for distributed base station processing. Unfortunately, the computational complexity of the BP algorithm grows exponentially with the number of interfering users at each base station. Reducing the computational load of marginalizing the a posteriori probabilities (APP) in BP nodes is challenging because this problem is generally rank deficient, and most well-known sub-optimal detection methods are not efficient in such conditions. In this paper we use an iterative group-wise multiuser detection algorithm to efficiently reduce the complexity of BP receiver. We evaluate the performance and complexity of the proposed RCBP algorithm under decomposed and clustered base station scenarios via simulations.