Multiway Distributed Wireless Relay Network With Projected Binary Quantization

Abstract

Wireless relay networking has been proposed as a solution to enable communication between disconnected users that otherwise are unable to reliably share information with one another due to distance or obstruction. Instead, a relay forwards a processed version of the data sent from one or more transmitters to one or more receivers. In this paper, we propose an energy efficient distributed relay networking structure where the spatial diversity of distributed multiple-input multiple-output systems is utilized. The distributed relay network consists of several single antenna relay nodes that are spatially distributed to receive transmissions from many users through an uplink. Each of the relay nodes performs a simple quantization operation to send a quantized version of the combination of the received signals through a downlink to a receiver (fusion center). Using the quantized signals broadcast by each relay node, we derive the maximum likelihood detector that can be implemented at each user to detect transmissions from other users. We also introduce suboptimal detectors that are less computationally complex than the maximum likelihood detector while having comparable performance. Analysis on the number of relay nodes required for the relay network to perform properly is also discussed. Our results are verified through simulations.