IDAF: Iterative Dual-Scale Attentional Fusion Network for Automatic Modulation Recognition
Recently, deep learning models have been widely applied to modulation recognition, and they have become a hot topic due to their excellent end-to-end learning capabilities. However, current methods are mostly based on uni-modal inputs, which suffer from incomplete information and local optimization....
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-09-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/23/19/8134 |
| _version_ | 1797575149588840448 |
|---|---|
| author | Bohan Liu Ruixing Ge Yuxuan Zhu Bolin Zhang Xiaokai Zhang Yanfei Bao |
| author_facet | Bohan Liu Ruixing Ge Yuxuan Zhu Bolin Zhang Xiaokai Zhang Yanfei Bao |
| author_sort | Bohan Liu |
| collection | DOAJ |
| description | Recently, deep learning models have been widely applied to modulation recognition, and they have become a hot topic due to their excellent end-to-end learning capabilities. However, current methods are mostly based on uni-modal inputs, which suffer from incomplete information and local optimization. To complement the advantages of different modalities, we focus on the multi-modal fusion method. Therefore, we introduce an iterative dual-scale attentional fusion (iDAF) method to integrate multimodal data. Firstly, two feature maps with different receptive field sizes are constructed using local and global embedding layers. Secondly, the feature inputs are iterated into the iterative dual-channel attention module (iDCAM), where the two branches capture the details of high-level features and the global weights of each modal channel, respectively. The iDAF not only extracts the recognition characteristics of each of the specific domains, but also complements the strengths of different modalities to obtain a fruitful view. Our iDAF achieves a recognition accuracy of 93.5% at 10 dB and 0.6232 at full signal-to-noise ratio (SNR). The comparative experiments and ablation studies effectively demonstrate the effectiveness and superiority of the iDAF. |
| first_indexed | 2024-03-10T21:35:38Z |
| format | Article |
| id | doaj.art-59f100e1ae084465ba91cf288408c556 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-10T21:35:38Z |
| publishDate | 2023-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-59f100e1ae084465ba91cf288408c5562023-11-19T15:03:13ZengMDPI AGSensors1424-82202023-09-012319813410.3390/s23198134IDAF: Iterative Dual-Scale Attentional Fusion Network for Automatic Modulation RecognitionBohan Liu0Ruixing Ge1Yuxuan Zhu2Bolin Zhang3Xiaokai Zhang4Yanfei Bao5Institute of Systems Engineering, Academy of Military Science of the People’s Liberation Army, Beijing 100083, ChinaInstitute of Systems Engineering, Academy of Military Science of the People’s Liberation Army, Beijing 100083, ChinaInstitute of Systems Engineering, Academy of Military Science of the People’s Liberation Army, Beijing 100083, ChinaNational Key Laboratory of Science and Technology on Communication, University of Electronic Science and Technology of China, Chengdu 611731, ChinaCollege of Communications and Engineering, Army Engineering University of PLA, Nanjing 210007, ChinaInstitute of Systems Engineering, Academy of Military Science of the People’s Liberation Army, Beijing 100083, ChinaRecently, deep learning models have been widely applied to modulation recognition, and they have become a hot topic due to their excellent end-to-end learning capabilities. However, current methods are mostly based on uni-modal inputs, which suffer from incomplete information and local optimization. To complement the advantages of different modalities, we focus on the multi-modal fusion method. Therefore, we introduce an iterative dual-scale attentional fusion (iDAF) method to integrate multimodal data. Firstly, two feature maps with different receptive field sizes are constructed using local and global embedding layers. Secondly, the feature inputs are iterated into the iterative dual-channel attention module (iDCAM), where the two branches capture the details of high-level features and the global weights of each modal channel, respectively. The iDAF not only extracts the recognition characteristics of each of the specific domains, but also complements the strengths of different modalities to obtain a fruitful view. Our iDAF achieves a recognition accuracy of 93.5% at 10 dB and 0.6232 at full signal-to-noise ratio (SNR). The comparative experiments and ablation studies effectively demonstrate the effectiveness and superiority of the iDAF.https://www.mdpi.com/1424-8220/23/19/8134automatic modulation recognitionmultimodal learningconvolutional neural networkattention mechanism |
| spellingShingle | Bohan Liu Ruixing Ge Yuxuan Zhu Bolin Zhang Xiaokai Zhang Yanfei Bao IDAF: Iterative Dual-Scale Attentional Fusion Network for Automatic Modulation Recognition Sensors automatic modulation recognition multimodal learning convolutional neural network attention mechanism |
| title | IDAF: Iterative Dual-Scale Attentional Fusion Network for Automatic Modulation Recognition |
| title_full | IDAF: Iterative Dual-Scale Attentional Fusion Network for Automatic Modulation Recognition |
| title_fullStr | IDAF: Iterative Dual-Scale Attentional Fusion Network for Automatic Modulation Recognition |
| title_full_unstemmed | IDAF: Iterative Dual-Scale Attentional Fusion Network for Automatic Modulation Recognition |
| title_short | IDAF: Iterative Dual-Scale Attentional Fusion Network for Automatic Modulation Recognition |
| title_sort | idaf iterative dual scale attentional fusion network for automatic modulation recognition |
| topic | automatic modulation recognition multimodal learning convolutional neural network attention mechanism |
| url | https://www.mdpi.com/1424-8220/23/19/8134 |
| work_keys_str_mv | AT bohanliu idafiterativedualscaleattentionalfusionnetworkforautomaticmodulationrecognition AT ruixingge idafiterativedualscaleattentionalfusionnetworkforautomaticmodulationrecognition AT yuxuanzhu idafiterativedualscaleattentionalfusionnetworkforautomaticmodulationrecognition AT bolinzhang idafiterativedualscaleattentionalfusionnetworkforautomaticmodulationrecognition AT xiaokaizhang idafiterativedualscaleattentionalfusionnetworkforautomaticmodulationrecognition AT yanfeibao idafiterativedualscaleattentionalfusionnetworkforautomaticmodulationrecognition |