.png){kind=icon}
Return
Deep Learning Based Physical-Layer Receiver With Pre-Denoise
.png){kind=icon}
OA
AI
DOI：10.1109/TVT.2023.3348803.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Traditional physical-layer receivers are typically composed of step-by-step serial processing modules heavily relying on complex hardware implementations and algorithms, where the optimal global performance cannot be guaranteed. In addition, system noise degrades the reliability of each module, resulting in the cumulative effect of errors. In this correspondence, we proposes a novel intelligent receiver consisting of a pre-denoised network and a stacked long-short term memory (LSTM) network to achieve more reliable information recovery. Specifically, the residual convolutional neural network (CNN) is employed to extract time-frequency characteristics of noise to denoise preprocessing, and then we utilize the stacked LSTM network to recover the information. Simulation results demonstrate that the proposed scheme can archive a bit error rate (BER) performance close to the ideal maximum likelihood algorithm and outperform the baselines.
Keywords:
Receivers
Wireless communication
Time-frequency analysis
Reliability
Convolutional neural networks
Network architecture
Convolution
Wireless communications
signal denoising
deep learning
convolutional neural network
stacked LSTM
Journal
IF:
7.1
Papers:
1.8W
Citations:
6.6W
Organization
View more organizations
Cited Papers
.png){kind=icon}
ShareSaveIntelligent and Reliable Deep Learning LSTM Neural Networks-Based OFDM-DCSK Demodulation Design
PREAI
ShareSaveUnsupervised Linear and Nonlinear Channel Equalization and Decoding Using Variational AutoencodersOAAI
OAAIShareSaveShareSaveShareSaveShareSave View more