Machine Learning–Based Adaptive Modulation and Coding Design

This chapter applies machine learning (ML) techniques to adaptive modulation and coding (AMC)‐aided wireless systems to allow them to adapt to the variations of channels. It introduces and analyzes two types of ML‐assisted AMC schemes in the context of the multiple‐input and multiple‐output and orthogonal frequency division multiplexing scenarios, where the conflicting demands for high date rate and high reliability are met by adjusting the modulation order and coding rate. Specifically, the chapter considers a supervised learning (SL) approach and a reinforcement learning approach. It first provides an overview of the ML‐assisted AMC, and of the AMC schemes specified in the IEEE 802.11n, and gives details of the modulation types and coding rates used in the standards. Then, it provides information on SL‐assisted AMC, where both the k‐nearest neighbor and support vector machine approaches are considered. Finally, the chapter considers corresponding simulation results and analyses.