基于GChebyNet-GRU的超宽带高精度室内定位方法  

UWB indoor positioning high⁃precision method based on GChebyNet⁃GRU

作　　者：许可 范馨月 XU Ke;FAN Xinyue(College of Mathematics and Statistics,Guizhou University,Guiyang 550025,China)

机构地区：[1]贵州大学数学与统计学院,贵州贵阳550025

出　　处：《电子设计工程》2024年第2期157-161,共5页Electronic Design Engineering

基　　金：贵州省数据驱动建模学习与优化创新团队(黔科合平台人才[2020]5016);贵州大学教改项目(XJG2021027);贵州大学一流课程培育项目(XJG2021040)。

摘　　要：超宽带高精度定位作为5G重大应用场景的关键技术,定位准确度极易受到室内复杂环境的干扰。为提高室内复杂环境下超宽带三维空间定位精度,提出了一种结合切比雪夫图卷积以及门控循环单元网络的定位模型。通过切比雪夫图卷积网络聚合静态的图结构和动态节点信息,结合门控循环网络捕捉节点间的依赖关系,进而精准预测靶点空间位置。实验结果表明,该模型相比于门控循环网络定位模型,无信号干扰情况下,X、Y、Z方向上的精度分别提高28.29%、41.48%、52.99%;有信号干扰情况下,精度分别提高34.95%、41.70%、36.43%,有效提高了基站处于信号干扰时室内定位的精度,更适合动态复杂的室内环境。As a key technology for major 5G application scenarios,UWB high⁃precision positioning is highly susceptible to interference from the complex indoor environment.To improve the accuracy of UWB 3D spatial localization in indoor complex environments,this paper proposes a localization model combining Chebyshev graph convolution and Gated Recurrent Unit.The static graph structure and dynamic node information are aggregated by the Chebyshev graph convolution network,and the depen⁃dencies between nodes are captured by the gated recurrent network to accurately predict the spatial location of the target.The experimental results show that compared with the gated recurrent network localization model,the accuracy in X,Y and Z directions is improved by 28.29%,41.48%and 52.99%,respectively,without signal interference,and 34.95%,41.70%and 36.43%,respectively,in the presence of signal interference.It is more suitable for dynamic and complex indoor environment.

关 键 词：超宽带 图结构 切比雪夫图卷积网络 门控循环单元 

分 类 号：TP212.9[自动化与计算机技术—检测技术与自动化装置] TN98[自动化与计算机技术—控制科学与工程]