神经网络提升低成本软件定义的无线电无人机探测平台定位准确率研究  

Research on Using Neural Network to ImproveAccuracy of Radio Platform for UAVDetection Defined by Low-Cost Software

作　　者：夏鸿[1] 何泉[1] 蒲磊 曾超 罗正华[2,3] XIA Hong;HE Quan;PU Lei;ZENG Chao;LUO Zhenghua(Jiuzhou Group Co.,Ltd.,Mianyang 621000,China;School of Electronical Information and Electronic Engineering,Chengdu University,Chengdu 610106,China;Sichuan Engineering Technology Research Centre of Time-Frequency Synchronization and Its Applications,Chengdu 610106,China)

机构地区：[1]四川九洲电器集团有限责任公司,四川绵阳621000 [2]成都大学电子信息与电气工程学院,四川成都610106 [3]四川省时频同步系统及应用工程技术研究中心,四川成都610106

出　　处：《成都大学学报（自然科学版）》2023年第3期262-267,282,共7页Journal of Chengdu University（Natural Science Edition）

基　　金：四川省科技计划重点研发项目(2021YFS0311、2021YFG0361)。

摘　　要：展示了用于比幅法测向的自研低成本软件定义无线电(software defined radio, SDR)系统,同时针对比幅法测向精度较低的问题,探讨了用来提升测向精度的神经网络,并重点验证了神经网络对自研系统接收信号的测向精度的提升程度.实验验证表明,对于自研系统接收信号在神经网络优化后的测向精度相较于直接通过比辐法解算接收信号的测向精度普遍有提高,其中性能最好的优化算法测向精度提高了近10%.因此,低成本硬件平台与神经网络形成了低成本且高精度的测向定位解决方案.In this paper,a low-cost software defined radio(SDR)system for amplitude comparison approach(ACA)method is demonstrated.To address the shortcoming of low accuracy of ACA,a neural network to improve the accuracy of the angle measurement is proposed.The amount of accuracy improvement by introducing neural network is mainly studied.Experimental results show that neural network has higher direction-finding accuracy than that by directly applying ACA method on received signals.The direction-finding accuracy of the best optimization algorithm increases approximately by 10%.The results of this paper show that combining the SDR system with the neural network can provide a low-cost and highly accurate solution to angle measurement problems.

关 键 词：比幅法 神经网络 低成本解决方案 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] V279[自动化与计算机技术—控制科学与工程]