| 摘要: |
| 非协作通信通用滤波多载波(Universal Filtered Multi-carrier，UFMC)信号子载波所存在的调制识别以及信噪比估计问题有待解决，但目前研究只针对于单一任务。对此，提出一种利用多任务学习框架的神经网络模型，同时解决调制识别以及信噪比估计任务。首先得到UFMC系统接收端信号，求解出信号同相正交分量作为输入特征；接着在多任务学习框架上构建神经网络，采用的神经网络是将卷积神经网络与长短时记忆网络串联；最后利用上述模型对两个任务进行联合求解。实验结果表明，所构建多任务学习模型性能优于单任务学习，在信噪比为0 dB时，子载波调制识别准确率提升7.71%，信噪比估计均方误差减小45.6%。 |
| 关键词: 通用滤波多载波（UFMC） 调制识别 信噪比估计 多任务学习 神经网络 |
| DOI：10.20079/j.issn.1001-893x.240421001 |
|
| 基金项目:重庆市自然科学基金项目（cstc2021jcyj-msxmX0836） |
|
| UFMC Modulation Recognition and SNR Estimation Based on Multi-task Learning |
| ZHANG Tianqi,WU Yunge,WU Xianyue,LI Chunyun |
| (School of Communication and Information Engineering,Chongqing University of Posts and Telecommunications,Chongqing 400065,China) |
| Abstract: |
| The modulation recognition and signal-to-noise ratio(SNR) estimation problems of the subcarriers of the universal filter multi-carrier(UFMC)signal in non-cooperative communication need to be solved,but the current research only focuses on a single task.Therefore,a neural network model using a multi-task learning framework is proposed to solve the modulation recognition and SNR estimation tasks at the same time.Firstly,the receiver signal of the UFMC system is obtained,and the orthogonal component of the signal is solved as the input feature.Then,a neural network is constructed on the multi-task learning framework.The neural network adopted is a convolution neural network and a long short-term memory network in series.Finally,the above model is used to solve the two tasks jointly.Experimental results show that the performance of the multi-task learning model constructed is better than that of single-task learning.When the SNR is 0 dB,the accuracy of subcarrier modulation recognition is improved by 7.71%,and the mean square error of SNR estimation is reduced by 45.6%. |
| Key words: universal filtered multi-carrier(UFMC) modulation identification SNR estimation multi-task learning neural network |
《电讯技术》编辑部 