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

An OFDM detection method based on deep learning to resist inter-subcarrier interference

A technology of inter-subcarrier interference and deep learning, which is applied in phase-modulated carrier systems, modulated carrier systems, baseband systems, etc., to achieve the effect of improving system performance

Active Publication Date: 2020-12-01
SOUTHEAST UNIV
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] Aiming at the ICI problem caused by Doppler frequency offset and phase noise in the OFDM system, the present invention provides an OFDM detection method based on deep learning for the OFDM communication system based on deep learning, and approximates the maximum by training the deep detection network. Excellent ML demodulator performance

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
  • An OFDM detection method based on deep learning to resist inter-subcarrier interference
  • An OFDM detection method based on deep learning to resist inter-subcarrier interference
  • An OFDM detection method based on deep learning to resist inter-subcarrier interference

Examples

Experimental program
Comparison scheme
Effect test

example 1

[0067] The simulation conditions are as follows:

[0068] Number of subcarriers 16 Deep Network Layers 30 Modulation QPSK batch size 500 sampling rate 1Mbps The number of channels in the training set 100000 Maximum Doppler Deviation 10kHz overall cycle times 500 Normalized Doppler Offset 0.16

[0069] The simulation channel uses the Rayleigh channel with six delay paths, and the relative power level uses the ITU Vehicle-A channel model parameters. figure 2 The simulation results under the above conditions are given. It can be seen that compared with the direct decorrelation method, the bit error rate of deep network detection is about 2dB lower.

example 2

[0071] The simulation conditions are as follows:

[0072] Number of subcarriers 16 Deep Network Layers 30 Modulation QPSK batch size 500 sampling rate 1Mbps The number of channels in the training set 100000 Maximum Doppler Deviation 15kHz overall cycle times 500 Normalized Doppler Offset 0.24

[0073] The simulation channel uses the Rayleigh channel with six delay paths, and the relative power level uses the ITU Vehicle-A channel model parameters. image 3 The simulation results under the above conditions are given. It can be seen that compared with the direct decorrelation method, the bit error rate of deep network detection is about 4dB lower. and figure 2 In comparison, it can also be seen that the performance of the deep detection network is basically not affected by the size of the Doppler frequency offset.

example 3

[0075] The simulation conditions are as follows:

[0076] Number of subcarriers 32 Deep Network Layers 30 Modulation QPSK batch size 500 sampling rate 1Mbps The number of channels in the training set 100000 Maximum Doppler Deviation 5kHz overall cycle times 500 Normalized Doppler Offset 0.16

[0077] The simulation channel uses Rayleigh channels with four delay paths, and the relative power level uses ITU Pedestrian-A channel model parameters. Figure 4 The simulation results under the above conditions are given. It can be seen that compared with the direct decorrelation method, the bit error rate of deep network detection is about 4-6dB lower.

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 an OFDM detection method based on deep learning to resist inter-subcarrier interference, which can be applied to high-speed mobile OFDM communication systems and OFDM systems with relatively large carrier phase noise in the millimeter wave band, and can effectively resist Doppler frequency offset and Inter-subcarrier interference caused by phase noise. In order to approach the ML detector, the present invention designs a deep network structure based on the projection gradient descent method by using the depth expansion method, and the training algorithm is the Adam algorithm, and adopts a micro-batch (batch) training method, and each batch contains multiple input and output OFDM symbols and the corresponding channel matrix H, that is, each batch reflects the change of the channel within a period of time. Through training, it traverses different channel information first, and then recycles these channel information for deep learning, so that the loss function converges to a small value. Using the trained deep detection network to demodulate OFDM signals can effectively improve the detection performance of OFDM systems affected by inter-subcarrier interference caused by large Doppler frequency offset or phase noise.

Description

technical field [0001] The invention relates to an OFDM detection method based on deep learning to resist inter-carrier interference (Inter-Carrier Interference, ICI), and belongs to the technical field of wireless mobile communication. Background technique [0002] OFDM technology can convert high-speed data streams into low-speed parallel data streams through serial-to-parallel conversion and modulate them on mutually orthogonal sub-carriers for transmission. Inter-symbol interference can be eliminated by adding a cyclic prefix. From the perspective of frequency domain, since the subcarrier spacing in OFDM is smaller than the coherent bandwidth of the channel, OFDM can resist frequency selective fading very well, and OFDM technology is easy to combine with multiple input and multiple output, using power allocation and adaptive modulation technology The spatial diversity and multiplexing gain can be obtained, and the requirements of channel transmission can be satisfied to ...

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
Patent Type & Authority Patents(China)
IPC IPC(8): H04L27/26H04L27/22H04L25/03G06N3/04
CPCH04L25/03821H04L27/22H04L27/265G06N3/045
Inventor 赵春明李骁敏姜明黄启圣
Owner SOUTHEAST 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