MCCFN

This repository contains the code for the paper:

A Multi-scale Complex-valued Convolutional Fusion Network for Automatic Modulation Recognition
Mobile Networks and Applications
DOI: 10.1007/s11036-026-02513-9
View-only full-text: https://rdcu.be/fBLi8

If you use this code, please cite the paper.

@article{Li2026,
  title = {A Multi-scale Complex-valued Convolutional Fusion Network for Automatic Modulation Recognition},
  ISSN = {1572-8153},
  url = {http://dx.doi.org/10.1007/s11036-026-02513-9},
  DOI = {10.1007/s11036-026-02513-9},
  journal = {Mobile Networks and Applications},
  publisher = {Springer Science and Business Media LLC},
  author = {Li, An and Li, Yue and Zhang, Qiang and Chen, Ping},
  year = {2026},
  month = Aug
}

Preparation

Data

Experiments were conducted on four datasets: RMLradio2016.10a, RMLradio2016.10b and RML22.

The dataset can be downloaded from the DeepSig official website.

Environment Setup

• pytorch = 2.4.0
• cuda = 11.8
• python = 3.8

Comparison with other models on the 2016a dataset.

Compare model:MCDformer,AWN,AvgNet,PETCGDNNand CLDNN

Some of the code is borrowed from MCDformer,AWN,CDSCNN and we thank them for their excellent work.

​

​