Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deep Learning-based Modulation Classification of Practical OFDM Signals for Spectrum Sensing

Byungjun Kim, Christoph Mecklenbräuker, Peter Gerstoft

TL;DR

A feature extraction algorithm characterizing the modulation of OFDM signals, which includes removing the effects of a synchronization error is proposed, which achieves a minimum accuracy of 97% accuracy with OTA data when SNR is above the value required for data transmission.

Abstract

In this study, the modulation of symbols on OFDM subcarriers is classified for transmissions following Wi-Fi~6 and 5G downlink specifications. First, our approach estimates the OFDM symbol duration and cyclic prefix length based on the cyclic autocorrelation function. We propose a feature extraction algorithm characterizing the modulation of OFDM signals, which includes removing the effects of a synchronization error. The obtained feature is converted into a 2D histogram of phase and amplitude and this histogram is taken as input to a convolutional neural network (CNN)-based classifier. The classifier does not require prior knowledge of protocol-specific information such as Wi-Fi preamble or resource allocation of 5G physical channels. The classifier's performance, evaluated using synthetic and real-world measured over-the-air (OTA) datasets, achieves a minimum accuracy of 97\% accuracy with OTA data when SNR is above the value required for data transmission.

Deep Learning-based Modulation Classification of Practical OFDM Signals for Spectrum Sensing

TL;DR

A feature extraction algorithm characterizing the modulation of OFDM signals, which includes removing the effects of a synchronization error is proposed, which achieves a minimum accuracy of 97% accuracy with OTA data when SNR is above the value required for data transmission.

Abstract

In this study, the modulation of symbols on OFDM subcarriers is classified for transmissions following Wi-Fi~6 and 5G downlink specifications. First, our approach estimates the OFDM symbol duration and cyclic prefix length based on the cyclic autocorrelation function. We propose a feature extraction algorithm characterizing the modulation of OFDM signals, which includes removing the effects of a synchronization error. The obtained feature is converted into a 2D histogram of phase and amplitude and this histogram is taken as input to a convolutional neural network (CNN)-based classifier. The classifier does not require prior knowledge of protocol-specific information such as Wi-Fi preamble or resource allocation of 5G physical channels. The classifier's performance, evaluated using synthetic and real-world measured over-the-air (OTA) datasets, achieves a minimum accuracy of 97\% accuracy with OTA data when SNR is above the value required for data transmission.
Paper Structure (16 sections, 15 equations, 12 figures, 7 tables, 1 algorithm)

This paper contains 16 sections, 15 equations, 12 figures, 7 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. System Objective
  3. Wi-Fi 6 PHY layer
  4. 5G DL PHY layer
  5. Proposed Algorithm
  6. OFDM parameter estimation
  7. Feature extraction
  8. Additional procedures for 5G signal
  9. Neural network classifier
  10. Evaluation
  11. Data collection
  12. Building classifier input
  13. Evaluation results
  14. AWGN channel data
  15. OTA data
  16. ...and 1 more sections

Figures (12)

  • Figure 1: (a) To capture DL Wi-Fi 6 and 5G signals and (b) Spectrum sensing scenario using USRP N310.
  • Figure 2: 5G Resource structure: (a) Resource grid and (b) Frame structure.
  • Figure 3: Flow chart of proposed modulation classification algorithm.
  • Figure 4: Flow chart of proposed classifier system: (a) Wi-Fi 6 and (b) 5G.
  • Figure 5: CNN-based modulation classifier structure. $\mathcal{N}_c$ is the number of modulations a classifier aims to recognize.
  • ...and 7 more figures