Distributed Optimization of Multiuser MIMO Relay Network Using Backpropagation Algorithm

• Published in: Conference record - Asilomar Conference on Signals, Systems, & Computers pp. 1205 - 1209
• Main Authors: Wang, Rui, Jiang, Yi
• Format: Conference Proceeding
• Language: English
• Published: IEEE 31.10.2021
• Subjects: Artificial neural networks; Backpropagation; backpropagation algorithm; Interference; interference suppression; nonlinear amplify-and-forward; Numerical simulation; quasi-neural network; relay network; Relay networks (telecommunication); Training; Transceivers
• DOI: 10.1109/IEEECONF53345.2021.9723157

This paper studies a multiuser multi-input multi-output (MIMO) relay network, where the transmit nodes and the relay nodes are subject to the nonlinear instantaneous power constraints. We introduce a perspective of regarding a relay network as a so-termed quasi-neural network by drawing its striking analogies to a (four-layer) artificial neural network (ANN) and propose a distributed scheme inspired by the backpropagation (BP) algorithm to optimize the transceivers. Using a crafty design of training sequences, the proposed scheme can be implemented with no channel state information (CSI) and no data exchange between the relay nodes. Numerical simulations verify the effectiveness of the proposed scheme. More important, it can suppress interferences from unknown directions.