End-to-End Learning for Uplink MU-SIMO Joint Transmitter and Non-Coherent Receiver Design in Fading Channels

In this paper, a novel end-to-end learning approach, namely JTRD-Net, is proposed for uplink multiuser single-input multiple-output (MU-SIMO) joint transmitter and non-coherent receiver design (JTRD) in fading channels. The basic idea lies the use of artificial neural networks (ANNs) to replace traditional communication modules at both sides. More specifically, side modeled as group parallel linear layers, which are responsible waveform design; formed by deep feed-forward network (DFNN) so provide detection (MUD) capabilities. entire JTRD-Net can be trained from end adapt channel statistics through learning. After training, work efficiently manner without requiring any levels state information (CSI). addition architecture, weight-initialization method, symmetrical-interval initialization, JTRD-Net. It shown that initialization outperforms conventional method (e.g. Xavier initialization) terms well-balanced convergence-rate among users. Simulation results show approach takes significant advantages reliability scalability over baseline schemes on i.i.d. complex Gaussian channels spatially-correlated