Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Channel state information reconstruction method based on deep learning

A technology of channel state information and deep learning, applied in channel estimation, baseband system, baseband system components, etc., can solve the problems of multiple iterations and increased computational complexity, and achieve the goal of reducing model overfitting and improving performance Effect

Active Publication Date: 2020-07-28
XI AN JIAOTONG UNIV
View PDF8 Cites 17 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Traditional reconstruction algorithms, such as Orthogonal Matching Pursuit (OMP), require a large number of iterations, which increases the computational complexity

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Channel state information reconstruction method based on deep learning
  • Channel state information reconstruction method based on deep learning
  • Channel state information reconstruction method based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0045] The present invention provides a channel state information reconstruction method based on deep learning. In a large-scale MIMO system using orthogonal frequency division multiplexing in FDD mode, the base station is set to N BS Antennas, K single-antenna users, and the number of subcarriers is N c . The schematic diagram of link transmission is as follows: figure 2 shown. The received signal on the nth subcarrier of the Kth user is expressed as:

[0046]

[0047] in, is the channel vector in the frequency domain, is the precoding vector, s kn ∈C is the data symbol, v kn ∈C is the additive noise on the nth subcarrier.

[0048] see image 3 , the present invention is a method for reconstructing channel state information based on deep learning, comprising the following steps:

[0049] S1. Obtain N in the frequency domain at the user end BS ×N c Dimensional channel matrix H, expressed as:

[0050]

[0051] in, means N BS ×N c A matrix whose dimensio...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a channel state information reconstruction method based on deep learning. The method comprises the following steps: obtaining an NBS * Nc-dimensional channel matrix H on a frequency domain at a user side; converting the obtained channel matrix H into an angle delay domain channel matrix Ha with only non-zero elements; in the angle time delay domain, acquiring a new dimension matrix by advancing in the intercepted channel matrix Ha, converting the new matrix into a non-sparse vector X with the dimension of 2N * 1, wherein the non-sparse vector X serves as to-be-compressed data, and carrying out compression through the compressed sensing technology to obtain to-be-fed back channel state information Y; reconstructing and training a channel state information network; and according to the trained ReNet network model, recovering Y as input data of the network and as output data of the network from the obtained channel state information Y to be fed back, and performinginverse Fourier transform to obtain original CSI after obtaining. The original CSI data obtained by the user side is processed, subsequent compression and feedback are facilitated, and recovery ofthe compressed data is completed through the network trained by the known samples.

Description

technical field [0001] The invention belongs to the technical field of communication, and in particular relates to a channel state information reconstruction method based on deep learning. Background technique [0002] Massive multiple-input multiple-output (Multiple-input Multiple-output, MIMO) system can improve the frequency spectrum and power utilization of wireless communication, and is one of the main technologies of the fifth generation wireless communication system. In a massive MIMO system, the base station usually needs to use channel state information (Channel State Information, CSI) for precoding, adaptive coding, user scheduling, etc. significant impact. [0003] In a Time Division Duplexing (TDD) system, due to the reciprocity between the uplink and downlink channels, the transmitter can estimate the CSI through the uplink channel, and then determine the CSI of the downlink channel through the reciprocity. However, in the Frequency Division Duplexing (FDD) mo...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): H04B7/06H04B7/0456H04L25/02G06N3/04
CPCH04B7/0663H04B7/0626H04B7/0456H04L25/0242H04L25/0204G06N3/045
Inventor 范建存梁培哲罗新民
Owner XI AN JIAOTONG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com