Symbol-Based Over-the-Air Digital Predistortion Using Reinforcement Learning

TL;DR

This paper introduces a reinforcement learning-based over-the-air digital predistortion method that optimizes transmitter performance without hardware or channel knowledge, improving symbol error rate and adjacent channel power ratio.

Contribution

It presents a novel symbol-based reinforcement learning algorithm for digital predistortion that operates over-the-air without requiring hardware or channel information.

Findings

Achieves lower symbol error rate compared to traditional methods.

Maintains satisfactory adjacent channel power ratio.

Operates effectively with generalized memory polynomial power amplifiers.

Abstract

We propose an over-the-air digital predistortion optimization algorithm using reinforcement learning. Based on a symbol-based criterion, the algorithm minimizes the errors between downsampled messages at the receiver side. The algorithm does not require any knowledge about the underlying hardware or channel. For a generalized memory polynomial power amplifier and additive white Gaussian noise channel, we show that the proposed algorithm achieves performance improvements in terms of symbol error rate compared with an indirect learning architecture even when the latter is coupled with a full sampling rate ADC in the feedback path. Furthermore, it maintains a satisfactory adjacent channel power ratio.