Method for improving deep learning channel estimation performance based on data augmentation of auto-encoder

Abstract

The invention discloses a method for improving deep learning channel estimation performance based on optimal data augmentation of an auto-encoder in a wireless communication scene, and the method comprises the following steps: building two basic convolutional neural network models for channel estimation; acquiring a training set of wireless communication channel estimation, and performing data augmentation based on an auto-encoder to obtain augmented data; estimating a basic convolutional neural network model by the augmented data through a channel to obtain a relationship between a mean square error value of the data augmentation method based on the auto-encoder on a test set and the size of a data set; based on a small amount of experimental data in the step 3, providing a simple straight line intersection point detection method, and obtaining the threshold value of the data set size when the self-encoder improves the performance of the wireless communication channel estimation model. According to the method, the data set threshold value when the self-encoder improves the performance of the wireless communication system to the greatest extent can be obtained, and the method has practical value for data augmentation of the model by using the self-encoder.

Description

technical field [0001] The invention relates to the technical field of wireless communication, in particular to a method for improving the performance of deep learning channel estimation based on autoencoder data augmentation. Background technique [0002] In communication systems, the received signal is often distorted by channel characteristics, which must be estimated and compensated at the receiver in order to recover the transmitted symbols. Typically, the receiver estimates the channel using some symbols called pilots, whose time-frequency locations are known to both the transmitter and receiver. Conventional pilot-based channel estimation methods include the least square method and the least mean square error method, both of which use pilot values ​​to estimate the unknown value of the channel response. How to reduce the complexity of these methods and ensure the estimation performance has always been a research hotspot in the field of wireless communication. Deep l...

AI Technical Summary

Problems solved by technology

However, the impact of data augmentation on deep learning-based channel estimation models has not been fully studied

[0004] Since the current typical data augmentation methods are designed for image classification tasks with semantic information, however, the channel estimation problem is usually a regression task, so typical data augmentation methods are not suitable for channel estimation problems

Images