Quantize-and-Forward Relay System with Autoencoder Using Multiple Antennas

In this paper, we propose a multi-input multi-output (MIMO) relay system with an autoencoder that jointly optimizes the transmitter and receiver applying deep learning. In this communication system, a memory-limited quantize-and-forward (QF) relay assists conventional point-to-point communications. With deep learning, the receiver (destination) does not need to estimate the channel information and avoids the high-complexity maximum-likelihood detection in the MIMO QF relay system, and thus this system can be a good alternative to next-generation communications which require high data rate and low latency.