The first part of the dataset HKDD_AMC36 of paper "Towards Next-Generation Signal Intelligence: A Hybrid Knowledge and Data-Driven Deep Learning Framework for Radio Signal Classification".

<strong>https://github.com/yexijoe/HKDD</strong> <br> <strong>@ARTICLE{10042021,</strong> <strong> author={Zheng, Shilian and Zhou, Xiaoyu and Zhang, Luxin and Qi, Peihan and Qiu, Kunfeng and Zhu, Jiawei and Yang, Xiaoniu},</strong> <strong> journal={IEEE Transactions on Cognitive Communications and Networking}, </strong> <strong> title={Towards Next-Generation Signal Intelligence: A Hybrid Knowledge and Data-Driven Deep Learning Framework for Radio Signal Classification}, </strong> <strong> year={2023},</strong> <strong> volume={},</strong> <strong> number={},</strong> <strong> pages={1-1},</strong> <strong> doi={10.1109/TCCN.2023.3243899}}</strong> <br> Here we publish the first part of the dataset HKDD_AMC36 used in the paper "<strong>Towards Next-Generation Signal Intelligence: A Hybrid Knowledge and Data-Driven Deep Learning Framework for Radio Signal Classification</strong>". Automatic modulation classification (AMC) can generally be divided into knowledge-based methods and data-driven methods. In this paper, we explore combining the knowledgebased method and data-driven technology to take full advantage of both and propose a hybrid knowledge and data-driven deep learning framework (HKDD) for AMC. To make the handcrafted features more discriminative, various traditional features are adopted, including instantaneous features, statistical features, and spectral features. In the HKDD framework, a feature fusion mechanism is proposed to integrate the features learned from the original signal with those processed by a fully connected network from the handcrafted features. Besides, an attention mechanism is implemented on the fused features to neglect immature features and highlight important features. To evaluate the performance of the proposed method, we construct two modulation classification datasets containing both traditional features and raw IQ data. Simulation results show that our proposed method has significant performance gain in both adequate-sample classification scenario and few-shot classification scenario.

https://github.com/yexijoe/HKDD
@ARTICLE{10042021,
作者：郑世廉、周小雨、张鲁新、戚佩涵、邱坤峰、朱佳伟、杨小牛，
期刊：《IEEE认知通信与网络汇刊》（IEEE Transactions on Cognitive Communications and Networking），
标题：《迈向下一代信号智能：面向无线电信号分类的知识与数据混合驱动深度学习框架》，
年份：2023，
卷：无，
期：无，
页码：1-1，
DOI：10.1109/TCCN.2023.3243899
}
本文发布了上述论文中所使用的数据集HKDD_AMC36的第一部分。自动调制分类（Automatic Modulation Classification, AMC）通常可分为基于知识的方法与数据驱动的方法两类。本文探索将基于知识的方法与数据驱动技术相结合，以充分发挥二者优势，并提出了一种面向AMC的混合知识与数据驱动深度学习框架（Hybrid Knowledge and Data-Driven Deep Learning Framework, HKDD）。为使人工设计特征更具区分度，本文采用了瞬时特征、统计特征与频谱特征等多种传统特征。在HKDD框架中，本文提出了一种特征融合机制，将从原始信号中学习得到的特征与通过全连接网络对人工设计特征处理后得到的特征进行融合。此外，本文还在融合特征上引入注意力机制，以过滤无效特征、突出关键特征。为验证所提方法的性能，本文构建了两份同时包含传统特征与原始IQ数据的调制分类数据集。仿真结果表明，所提方法在充足样本分类场景与少样本分类场景下均取得了显著的性能提升。