Performance Analysis of Multi-Way Quantized Distributed Relay Networking

Wireless relay networking has been proposed as a solution for extending coverage in the past few decades. The relay network facilitates communication between users who are unable to reliably share information due to severe pathloss or blockage. In this paper, we utilize spatial diversity of distributed multiple-input multiple-output systems for the relay network. Hence, the relay network consists of many separate relay nodes. Due to limited computational power, we assume each relay node receives the sum of the transmissions from all users and then performs one-bit quantization. The quantization bits from the relay network are broadcast back to the users through a downlink channel that is modeled as a low-rate binary symmetric channel. Based on the noisy quantization bits from the relay network and its own prior transmission, each user detects the transmitted symbols from other users. We first derive the maximum likelihood detector for the described system. Then we develop a sub-optimal detector called the orthogonal subset maximum likelihood (OSML) detector, which exploits only a subset of relay nodes for detection, to reduce the computational complexity. By using combinatorial geometry, we derive the minimum number of required relay nodes for the OSML detector to operate. The derived results are verified through numerical simulations.