Diffusion SigFormer for Interference Time-series Signal Recognition

Zhouhuaji, Xiaoyu Hao, Xu Liu, Lingling Li, Fang Liu, Licheng Jiao

Keywords: Anti-interference electromagnetic signal recognition, diffusion, SigFormer, modulation, bluetooth

Abstract:

The various interferences in the actual environment make electromagnetic signal recognition challenging, and this topic has extremely important application value.
In this paper, a novel interference signal recognition transformer is proposed, named Diffusion SigFormer.
Firstly, we explored the interference law of electromagnetic signals and designed a signal interference mechanism.
Secondly, diffusion signal denoising modulewas proposed to denoise the input interference signal. We also use various types of noise to improve its denoising effect on electromagnetic signals.
Thirdly, SigFormer is designed to extract and classify the denoised signal.
For the characteristics of electromagnetic signals, SigFormer leverages 1-D Patch Embedding and combines transformer with convolution.
Finally, we conducted experimental verification on datasets RML2016.10a, RML2016.10b and BT dataset.
The experimental results show that the proposed method has excellent anti-interference ability.

Primary Area: learning on time series and dynamical systems

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 9606