|
|
基于多尺度特征融合的轻量级自动调制识别网络
|
Abstract:
针对现有自动调制识别模型在资源受限场景下部署困难的问题，本文建立了基于多尺度特征融合的轻量级自动调制识别网络，结合多尺度深度可分离卷积、通道注意力机制与循环神经网络的高效识别架构，采用端到端的单输入设计，通过多分支深度可分离卷积提取信号的时频特征，并利用LSTM单元捕捉时序依赖关系，实现了轻量准确的信号调制方式识别功能。实验表明，所提模型在保持高识别精度的同时，显著降低了计算复杂度与内存占用，适用于边缘计算设备部署。
To address the deployment challenges of existing automatic modulation recognition (AMR) models in resource-limited scenarios, this paper introduces a lightweight network based on multi-scale feature fusion. The proposed architecture efficiently integrates multi-scale depthwise separable convolution, a channel attention mechanism, and a recurrent neural network. It employs an end-to-end, single-input design. This design uses multi-branch depthwise separable convolutions to extract time-frequency features and LSTM units to capture temporal dependencies. This enables accurate and lightweight modulation recognition. Experimental results show that our model maintains high recognition accuracy while significantly reducing computational complexity and memory usage. It is, therefore, suitable for deployment on edge computing devices.
| [1] | 濮希同. 面向调制识别的深度学习模型轻量化关键技术研究[D]: [硕士学位论文]. 成都: 电子科技大学, 2025. |
| [2] | O’Shea, T.J., Corgan, J. and Clancy, T.C. (2016) Convolutional Radio Modulation Recognition Networks. In: Jayne, C. and Iliadis, L., Eds., Engineering Applications of Neural Networks, Springer International Publishing, 213-226. https://doi.org/10.1007/978-3-319-44188-7_16 |
| [3] | Xu, J., Luo, C., Parr, G. and Luo, Y. (2020) A Spatiotemporal Multi-Channel Learning Framework for Automatic Modulation Recognition. IEEE Wireless Communications Letters, 9, 1629-1632. https://doi.org/10.1109/lwc.2020.2999453 |
| [4] | Liu, K.T. and Wu, S. (2024) Automatic Modulation Classification Based on Time-Attention Mechanism and LSTM Neural Networks. 2024 7th International Conference on Information Communication and Signal Processing (ICICSP), Zhoushan, 21-23 September 2024, 112-117. https://doi.org/10.1109/icicsp62589.2024.10809243 |
| [5] | 张富鑫. 基于深度学习的自动调制识别技术研究[D]: [硕士学位论文]. 成都: 电子科技大学, 2023. |
| [6] | Zhu, Z., Zhou, N., Wang, Z., Liang, J. and Wang, Z. (2024) MSGNet: A Multi-Feature Lightweight Learning Network for Automatic Modulation Recognition. IEEE Communications Letters, 28, 2553-2557. https://doi.org/10.1109/lcomm.2024.3414617 |
| [7] | Zeng, D., Xiao, Y., Liu, W., Du, H., Zhang, E., Zhang, D., et al. (2025) Efficient Automatic Modulation Classification in Nonterrestrial Networks with Snn-Based Transformer. IEEE Internet of Things Journal, 12, 2562-2573. https://doi.org/10.1109/jiot.2024.3467284 |
| [8] | Tekbiyik, K., Ekti, A.R., Gorcin, A., Kurt, G.K. and Kececi, C. (2020) Robust and Fast Automatic Modulation Classification with CNN under Multipath Fading Channels. 2020 IEEE 91st Vehicular Technology Conference (VTC2020-Spring), Antwerp, 25-28 May 2020, 1-6. https://doi.org/10.1109/vtc2020-spring48590.2020.9128408 |
| [9] | Rajendran, S., Meert, W., Giustiniano, D., Lenders, V. and Pollin, S. (2018) Deep Learning Models for Wireless Signal Classification with Distributed Low-Cost Spectrum Sensors. IEEE Transactions on Cognitive Communications and Networking, 4, 433-445. https://doi.org/10.1109/tccn.2018.2835460 |
| [10] | Hermawan, A.P., Ginanjar, R.R., Kim, D. and Lee, J. (2020) CNN-Based Automatic Modulation Classification for beyond 5G Communications. IEEE Communications Letters, 24, 1038-1041. https://doi.org/10.1109/lcomm.2020.2970922 |
