A Deep Learning Receiver for Non-Linear Transmitter

Non-linearity of wireless transceivers, specifically power amplifier (PA) non-linearity, could pose major limitations towards having high throughput, and cost energy efficient communication systems. Such from the PA is typically compensated in transmitter, e.g. by applying back-off or performing digital-pre-distortion (DPD) aiming to linearize transmitter. However, leads lower efficiency, output power, hence coverage; DPD results higher complexity transmitters. This paper presents an alternative approach based on a receiver method perform signal detection presence distortions due non-linearity. We propose technique using artificial neural networks (ANN) compensate for non-linearity at side. The link-level simulation pre-trained network models synthesized training data. confirm that designed can tolerate distortions, allow be reduced, leading improving coverage, spectral throughput.