摘要
为了实现低成本、小型化无线电监测与识别,基于嵌入式GPU设计了一种通信信号调制方式识别系统。该系统由嵌入式GPU计算单元、无线收发单元、无线通信与定位单元等组成,具有通信信号采集、智能处理、信息远程发布等功能。在该系统上部署了通信信号采集、二维时频图计算和深度学习分类网络计算等程序。经过大量实际测试,针对BPSK、QPSK、8-PSK等11种调制信号,平均识别精度为89.2%。
In order to realize radio monitoring and recognition with low-cost and miniaturization,a communication signal modulation recognition system based on embedded GPU is designed.The system mainly consists of four units including embedded GPU computing unit,wireless transceiver unit,wireless communication unit and positioning unit,and it has the functions of communication signal acquisition,intelligent processing and information remote release.Communication signal acquisition,two-dimensional time-frequency diagram calculation and deep learning classification network calculation are deployed on the system.The average recognition accuracy is 89.2%by testing on eleven modulation signals such as BPSK,QPSK,8-PSK and so on.
作者
曹洁
宋蓓蓓
CAO Jie;SONG Bei-bei(Xi'an Tonghe Telecom Equipment Testing Co.,Ltd,Xi'an 710061,China;Department of Communication Engineering,School of Information Engineering,Chang'an University,Xi'an 710064,China)
出处
《无线通信技术》
2020年第3期21-25,共5页
Wireless Communication Technology
关键词
调制识别
嵌入式GPU
深度学习
modulation recognition
embedded GPU
deep learning
作者简介
曹洁,1982年生,女,汉族,工程师,主要从事无线通信设备检测技术的研究。
