Neural Network–Based Wireless Channel Prediction

Abstract

This chapter provides a comprehensive introduction to channel prediction methods with an emphasis on neural network-based prediction. It first briefly describes adaptive transmission systems using transmit antenna selection and opportunistic relaying as examples, followed by the impact of outdated channel state information (CSI) on the performance of adaptive transmission systems. The chapter provides a mathematical model to quantify the inaccuracy of outdated CSI and then uses the opportunistic relay selection system as an example to illustrate the impact of outdated CSI on the performance of adaptive transmission systems. It then reviews two kinds of classical prediction methods: parametric and autoregressive models. The chapter also details the principles of recurrent neural network-based predictors applied from flat-fading single-antenna channels to frequency-selective multi-antenna channels, as well as their achievable performance and computational complexity.