Abstract:Recently, machine learning-based channel estimation has attracted much attention. The performance of machine learning-based estimation has been validated by simulation experiments. However, little attention has been paid to the theoretical performance analysis. In this paper, we investigate the mean square error (MSE) performance of machine learning-based estimation. Hypothesis testing is employed to analyze its MSE upper bound. Furthermore, we build a statistical model for hypothesis testing, which holds when the linear learning module with a low input dimension is used in machine learning-based channel estimation, and derive a clear analytical relation between the size of the training data and performance. Then, we simulate the machine learning-based channel estimation in orthogonal frequency division multiplexing (OFDM) systems to verify our analysis results. Finally, the design considerations for the situation where only limited training data is available are discussed. In this situation, our analysis results can be applied to assess the performance and support the design of machine learning-based channel estimation.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the insufficient performance analysis of machine learning (ML) in channel estimation. Although machine learning methods have shown performance close to the optimal estimation (i.e., the minimum mean - square error estimation, MMSE) in channel estimation, there is currently a lack of theoretical analysis of this performance. Specifically, the paper focuses on: 1. **Theoretical analysis of the mean - square error (MSE) difference**: The paper analyzes the MSE difference between the machine - learning - based channel estimation and the optimal channel estimation (MMSE) through the method of hypothesis testing. The research shows that even if the input data is not in the training set, the MSE difference has an upper bound, which provides theoretical support for the machine - learning - based channel estimation. 2. **The relationship between the size of the training data set and performance**: The paper further specifies the distribution function in the hypothesis testing and derives the analytical relationship between the size of the training data set and performance. This relationship is demonstrated through numerical results, and the sufficient size of the training data set is determined. 3. **Design considerations in the case of limited training data**: The paper discusses how to design the machine - learning - based channel estimation when only limited training data is available. In this case, a linear learning module may be adopted, and the analysis results of the paper can be used to predict performance and help determine the input dimension of the learning module. In summary, the main objective of the paper is to fill the gap in the performance analysis of the machine - learning - based channel estimation through theoretical analysis, especially to conduct in - depth discussions on the MSE difference and the influence of the size of the training data set. These analysis results are not only helpful for understanding the effectiveness of the machine - learning - based channel estimation but also provide guidance for the design in practical applications.

Performance Analysis on Machine Learning-Based Channel Estimation

Optimal-tuned Weighted LASSO for Massive MIMO Channel Estimation with Limited RF Chains

Channel Estimation Algorithm for Mobile OFDM Receivers

Machine Learning-Based 5G-and-Beyond Channel Estimation for MIMO-OFDM Communication Systems

Performance Analysis of Channel Estimation Based on Base Extended Model under Fast Time-Varying Channels

Knowledge-Driven Machine Learning-based Channel Estimation in Massive MIMO System

A LSE and Sparse Message Passing-Based Channel Estimation for Mmwave MIMO Systems.

Knowledge-Driven Machine Learning: Concept, Model and Case Study on Channel Estimation

Deep Learning for Channel Estimation: Interpretation, Performance, and Comparison

Machine Learning for Wireless Communication Channel Modeling: An Overview

A Machine Learning Based Channel Parameter Prediction Method Utilizing Ray Tracing Simulation Dataset

Machine learning based low-complexity channel state information estimation

Channel Estimation Aware Performance Analysis for Massive MIMO With Rician Fading

Performance Analysis of TDD System in Channel Estimate

Dual-Diagonal LMMSE Channel Estimation for OFDM Systems

Deep Learning for MIMO Channel Estimation: Interpretation, Performance, and Comparison

Blind Performance Prediction for Deep Learning Based Ultra-Massive MIMO Channel Estimation

MEP-Based Channel Estimation under Complex Communication Environment

Performance Evaluation of Offset Spatial Modulation Systems with Practical Channel Estimation

Performance Evaluation of an OFDM-based LMDS Using Measured Channel Models.

Power of Deep Learning for Channel Estimation and Signal Detection in OFDM Systems