• Home
  • Search
  • Specific Emitter Identification Method for Limited Samples via Time\u2013Wavelet Spectrum Consistency
  • Cite Icon3
  • https://doi.org/10.3390/s25030648Copy DOI Icon

Specific Emitter Identification Method for Limited Samples via Time\u2013Wavelet Spectrum Consistency

Show More
  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Specific emitter identification (SEI) is a technique that identifies the emitter through physical layer features contained in radio signals, and it is widely used in tasks such as identifying illegal transmitters and authentication. Thanks to the development of deep learning, SEI tasks based on deep learning have achieved significant improvements in recognition performance. However, illegal transmitters often broadcast for short durations and at low frequencies, resulting in very limited available training samples. In such cases, directly training deep learning models may lead to underfitting issues, thereby reducing recognition accuracy. In this paper, to address the issue of traditional methods struggling to classify transmitters when there is a scarcity of emitter samples and limited training data, we propose a method based on TFC-CNN. We propose a TFC-CNN method. Specifically, we first use continuous wavelet transform (CWT) as a data augmentation method to construct time–wavelet spectrum sample pairs. Then, we use complex-valued neural networks (CVNNs) and deep convolutional neural networks (DCNNs) to extract the hidden emitter identity features from the time and wavelet spectrum samples. We train the model using the normalized temperature-scaled cross-entropy (NT-Xent) loss and cross-entropy (CE) loss, ensuring the consistency of feature vector distributions across the two modalities with cosine loss. Finally, we fine-tune the model to achieve SEI tasks with few samples. Experimental results on open-source WiFi datasets and automatic dependent surveillance–broadcast (ADS-B) datasets show that our proposed method outperforms existing state-of-the-art methods. With only 5% of the training samples, the recognition accuracy for the ADS-B test dataset is 84.10%, and for the WiFi test dataset, it is 96.99%.

Similar Papers
  • Research Article
  • Citations138

Few-Shot Specific Emitter Identification via Deep Metric Ensemble Learning

  • Dec 15, 2022
  • IEEE Internet of Things Journal
  • Yu Wang +5
  • Research Article
  • Citations22

Specific emitter identification of radar based on one dimensional convolution neural network

  • May 01, 2020
  • Journal of Physics: Conference Series
  • Yao Xiao +1
  • Research Article
  • Citations284

An Efficient Specific Emitter Identification Method Based on Complex-Valued Neural Networks and Network Compression

  • Aug 01, 2021
  • IEEE Journal on Selected Areas in Communications
  • Yu Wang +5
  • PDF
  • Research Article
  • Citations13

Transfer Learning-Based Specific Emitter Identification for ADS-B over Satellite System

  • Jun 07, 2024
  • Remote Sensing
  • Mingqian Liu +4
  • Research Article
  • Citations9

Enhancing Specific Emitter Identification: A Semi-Supervised Approach With Deep Cloud and Broad Edge Integration

  • Jan 01, 2025
  • IEEE Transactions on Information Forensics and Security
  • Yibin Zhang +5
  • Research Article

An Adaptive Deep Ensemble Learning for Specific Emitter Identification.

  • Oct 09, 2025
  • Sensors (Basel, Switzerland)
  • Peng Shang +5
  • Conference Article
  • Citations7

Attention-CNN-Aided Specific Emitter Identification Method with Limited Radio Frequency Dataset

  • Nov 11, 2022
  • Shufei Wang +4
  • Research Article
  • Citations117

Semi-Supervised Specific Emitter Identification Method Using Metric-Adversarial Training

  • Jun 15, 2023
  • IEEE Internet of Things Journal
  • Xue Fu +6
  • Research Article
  • Citations29

Improving RF-DNA Fingerprinting Performance in an Indoor Multipath Environment Using Semi-Supervised Learning

  • Jan 01, 2024
  • IEEE Transactions on Information Forensics and Security
  • Mohamed K M Fadul +5
  • Research Article
  • Citations3

Resource-Constrained Specific Emitter Identification Based on Efficient Design and Network Compression.

  • Apr 04, 2025
  • Sensors (Basel, Switzerland)
  • Mengtao Wang +3
  • Research Article
  • Citations52

Semi-Supervised Specific Emitter Identification via Dual Consistency Regularization

  • Nov 01, 2023
  • IEEE Internet of Things Journal
  • Xue Fu +6
  • Conference Article
  • Citations16

A method for specific emitter identification based on surrounding-line bispectrum and convolutional neural network

  • Nov 20, 2020
  • Hao Ji +3
  • Research Article
  • Citations22

A probe-feature for specific emitter identification using axiom-based grad-CAM

  • Jul 10, 2022
  • Signal Processing
  • Mingzhe Zhu +5
  • Conference Article
  • Citations4

ADS-B Message Authentication Using Features of Signal in Transition Regions

  • Dec 01, 2019
  • Nan Jiang +4
  • Research Article
  • Citations47

Multi-feature fusion for specific emitter identification via deep ensemble learning

  • Dec 28, 2020
  • Digital Signal Processing
  • Zhang-Meng Liu
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.