| 330 | 6 | 141 |
| 下载次数 | 被引频次 | 阅读次数 |
针对模拟、数字调制方式在简单网络下识别率不高的问题,提出了一种基于特征融合、自注意力机制、并联神经网络的调制识别算法——MACLNN。使用8个统计量特征参数组合和IQ数据分别作为卷积神经网络(Convolutional Neural Network, CNN)和卷积长短时神经网络的输入,由自注意力机制重新分配特征的权重,再通过并联层进行特征融合,最终完成11类调制方式的识别。仿真结果表明,在高信噪比下识别准确率可达到94.1%,使用复杂度较低的模型获得了高于同类算法的识别精度。
Abstract：To solve the problem of low recognition rate of analog and digital modulation methods in simple networks,a modulation recognition algorithm——MACLNN based on feature fusion,self-attention mechanism and parallel neural network is proposed. Using eight statistical feature parameter combinations and IQ data as the input of the convolutional neural network and the convolutional long and short-term neural network respectively,the weight of the features is redistributed by the self-attention mechanism,and then the feature fusion is performed through the parallel layers,and finally recognitions of eleven types of modulation methods are completed. The simulation results show that the recognition accuracy can reach 94. 1% under high signal-to-noise ratio,which is higher than that of similar algorithms and has relatively few training parameters of the model.
[1]WANG F,DOBRE O A,CHAN C,et al.Fold-based Kolmogorov-smirnov Modulation Classifier[J].IEEE Signal Processing Letters,2016,23(7):1003-1007.
[2]DOBRE O A.Signal Identification for Emerging Intelligent Radios:Classical Problems and New Challenges[J].IEEEInstrumentation&Measurement Magazine,2015,18(2):11-18.
[3]ZHANG M,DIAO M,GUO L.Convolutional Neural Networks for Automatic Cognitive Radio Waveform Recognition[J].IEEE Access,2017,5:11074-11082.
[4]TANG B,TU Y,ZHANG Z,et al.Digital Signal Modulation of Classification with Data Augmentation Using Generative Adversarial Nets in Cognitive Radio Networks[J].IEEE Access,2018,6:15713-15722.
[5]戴江安,栾声扬,赵明龙,等.脉冲噪声下基于平滑循环相关熵谱的调制识别方法[J].通信学报,2021,42(12):121-133.
[6]张艳秋.基于高阶累积量的数字信号调制识别[J].信息通信,2016(2):27-30.
[7]张利,李青.基于高阶累积量的调制识别算法的研究[J].信息工程大学学报,2017,18(4):403-408.
[8]杨杰,刘珩,卜祥元,等.通信信号调制识别:原理与算法[M].北京:人民邮电出版社,2014.
[9]HOCHREITER S,SCHMIDHUBER J.Long Short-Term Memory[J].Neural Computation,1997,9(8):1735-1780.
[10]O′SHEA T J,WEST N.Radio Machine Learning Dataset Generation with GNU Radio[C]∥Proceedings of the 6th GNU Radio Conference.Huntsville:[s.n.],2016:69-74.
[11]谭继远,张立民,钟兆根.基于卷积双向长短期神经网络的调制方式识别[J].火力与指挥控制,2020,45(6):129-134.
[12]JIANG J,WANG Z,ZHAO H,et al.Modulation Recognition Method of Satellite Communication Based on CLDNNModel[C]∥2021 IEEE 30th International Symposium on Industrial Electronics (ISIE).Kuoto:IEEE,2021:1-6.
[13]赵梓存,栾风虎,禹永植.一种基于CLSNN网络的通信信号自动调制识别算法[J].应用科技,2022,49(3):50-54.
[14]ZHANG Z F,LUO H,WANG C,et al.Automatic Modulation Classification Using CNN-LSTM Based Dual-stream Structure[J].IEEE Transactions on Vehicular Technology,2020,69(11):13521-13531.
[15]XIONG X J,WU H L,WU R J,et al.Stacked Hourglass Network with Integration of Channel Information for Automatic Modulation Recognition[J].Signal,Image and Video Processing,2022(prepublish):1-9.
基本信息:
中图分类号:TN911.3
引用信息:
[1]吴美霖,高瑜翔,涂雅培,等.基于特征融合和MACLNN的通信信号自动调制识别[J].无线电工程,2022,52(11):1970-1976.