多特征融合的半监督流形约束定位方法  

Semi-supervised Manifold Constraint Localization Method with Multi-feature Fusion

作　　者：钱政 严亮 孙顺远[1] QIAN Zheng;YAN Liang;SUN Shunyuan(Key Laboratory of Advanced Process Control for Light Industry of Ministry of Education,School of Internet of Things Engineering,Jiangnan University,Wuxi 214122,Jiangsu Province,China)

机构地区：[1]江南大学物联网工程学院,轻工先进控制教育部重点实验室,江苏无锡214122

出　　处：《吉林大学学报（理学版）》2024年第5期1219-1227,共9页Journal of Jilin University:Science Edition

基　　金：国家自然科学基金(批准号:61702228);江苏省自然科学基金(批准号:BK20170198)。

摘　　要：针对无线保真(WiFi)和低功耗蓝牙(BLE)指纹定位方法需要大量标记训练样本以及单模定位精度和稳定性难以满足大规模定位场景需求的问题,提出一种融合WiFi和BLE信号的半监督流形约束定位方法.实验结果表明:该方法与单一特征相比,每一维度归一化方差稳定在0.08以下,定位精度约提高25个百分点;使用分别构建流形约束的半监督学习方法时,能使定位过程中所需标记样本数量减少约90%.因此,该方法能极大减少需标记的样本数量,并有效提高定位的稳定性和精度.Aiming at the problems that wireless fidelity(WiFi)and bluetooth low energy(BLE)fingerprint localization methods required a large number of labeled training samples and that the accuracy and stability of single-mode localization were difficult to meet the requirements of large-scale localization scenarios,we proposed a semi-supervised manifold constraint localization method that fused WiFi and BLE signals.The experimental results show that compared with a single feature,the no rmalized variance of each dimension of the proposed method is stable below 0.08,and the accuracy of localization is improved by about 25 percentage points.When the semi-supervised learning method is used to construct manifold constraints separately,the number of labeled samples required in the localization proce ss can be reduced by about 90%.Therefore,this method can greatly reduce the number of required label samples,and effectively improve the stability and accuracy of localization.

关 键 词：多特征融合 半监督学习 流形正则化 无线保真(WiFi) 低功耗蓝牙 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]