Channel estimation method based on compressed sensing and deep learning, medium and equipment

Abstract

The invention discloses a channel estimation method based on compressed sensing and deep learning, a medium and equipment. The method comprises the steps that a base station end of an orthogonal frequency division multiplexing system sends a signal to a user end in a downlink in a comb-shaped pilot frequency form; the user side in the downlink obtains the sent pilot frequency receiving signal y and feeds back the signal y to the base station side; the base station end of the orthogonal frequency division multiplexing system performs channel estimation based on an ASJOMP algorithm of compressed sensing according to the obtained receiving pilot signal y by using the structured sparse characteristic of a time delay domain sparse channel to obtain an initial estimation channel, builds a noise reduction neural network based on deep learning, and trains a DnNet network by using an existing sample to obtain a network parameter theta; and de-noising the obtained initial estimation channel according to the DnNet network to obtain finally estimated channel state information. A lightweight network is adopted, so that the training process is faster, the calculated amount is reduced, and finally, an accurate CSI estimation value is obtained.

Description

technical field [0001] The invention belongs to the field of communication technology, and in particular relates to a channel estimation method, medium and equipment based on compressed sensing and deep learning. Background technique [0002] In a Frequency Division Duplexing (FDD) system, since the channel does not have reciprocity, downlink channel information cannot be directly obtained through uplink estimation. In order to obtain downlink channel state information (channel state information, CSI), downlink pilot training must be performed on the downlink channel. In this scenario, the base station sends a pilot to the user, and the user estimates channel information after receiving the pilot and feeds it back to the base station, or the user directly feeds back the received pilot information to the base station, and the base station jointly performs channel estimation. In this case, in order to obtain more accurate channel information, traditional channel information a...

AI Technical Summary

Problems solved by technology

[0004] Traditional methods based on compressed sensing, such as Orthogonal Matching Pursuit (OMP) and Compressive Sampling Matching Pursuit (CoSaMP), require prior knowledge of the known channel, such as channel sparsity, etc. , while the estimated channel still contains a lot of noise

Images