Scholarlink

This paper suggests a deep-learning approach for radio signal denoising using a Convolutional Autoencoder. It provides a basic understanding of machine learning and how it can mitigate the unnecessary noise from a radio signal. It can be trained specifically for a single kind of signal or combined can be used as a component of a denoising module in a pipeline. We have conducted tests and training on artificial radio data generated using mathematical functions and introduced different noise functions to mimic real world scenarios. Artificial radio data provides the freedom of flexibility to change different signal parameters and experiment on it using an autoencoder. Autoencoder can further be used to classify noise, protocol and modulation to improve the accuracy for denoising.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Noise Mitigation in Radio Signals Using Convolutional Neural Network Autoencoders

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Shivam Pathak,
  • Er. Parul Awasthi,
  • Er. Amit Katiyar,
  • Nikhil Kumar Singh,
  • Paritosh Narayan Pandey,
  • Priyanshu Tiwari,
  • Sushil Kumar Yadav

摘要

This paper suggests a deep-learning approach for radio signal denoising using a Convolutional Autoencoder. It provides a basic understanding of machine learning and how it can mitigate the unnecessary noise from a radio signal. It can be trained specifically for a single kind of signal or combined can be used as a component of a denoising module in a pipeline. We have conducted tests and training on artificial radio data generated using mathematical functions and introduced different noise functions to mimic real world scenarios. Artificial radio data provides the freedom of flexibility to change different signal parameters and experiment on it using an autoencoder. Autoencoder can further be used to classify noise, protocol and modulation to improve the accuracy for denoising.