基于差分星座轨迹图的多任务 802. 11b / g 信号识别方法. (Chinese)

For the modulation recognition and emitter individual recognition of 802. 11b / g wireless signal, a multi-task convolution neural network recognition method based on differential constellation trace figure is presented. The modulation recognition and emitter individual recognition are regarded as two interrelated learning tasks. The characteristics of the differential constellation trace figure are extracted through the depth network with shared parameters, and classified and recognized by two branch networks with different structures. At the same time, the two tasks are jointly optimized and mutually promoted to learn. In the experiment, six different routers are used for verification. The results show that compared with the recognition method of single task model, the multi-task model takes less training time and memory than the sum of the two single task models. Meanwhile, the recognition rate of individual emitter and modulation mode is improved by 1. 17% and 3% on average respectively. [ABSTRACT FROM AUTHOR]

针对 802. 11b / g 无线信号的调制方式识别和辐射源个体识别问题,提出了一种基于差分星 座轨迹图的多任务卷积神经网络识别方法. 将调制识别和辐射源个体识别看作两个相互关联的学 习任务,通过共享参数的深度网络提取差分星座轨迹图的特征,并由结构不同的两个分支网络进行 分类识别,同时对这两个任务进行联合优化训练并相互促进学习. 实验中使用 6 个不同的路由器进 行验证,结果表明相比于单任务模型的识别方法,多任务模型所用的训练时长和模型所占内存均比 两个单任务模型之和少,同时对辐射源个体, 调制方式的识别率分别平均提高了 1. 17% 和 3%. [ABSTRACT FROM AUTHOR]

Copyright of Telecommunication Engineering is the property of Telecommunication Engineering and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)