基于EKF融合的室内定位技术研究  

Research on indoor positioning technology based on EKF fusion

作　　者：史明泉[1] 李妮芝 崔丽珍[1] 秦岭 SHI Mingquan;LI Nizhi;CUI Lizhen;QIN Ling(School of Information Engineering,Inner Mongolia University of Science and Technology,Baotou 014010,China)

机构地区：[1]内蒙古科技大学信息工程学院,内蒙古包头014010

出　　处：《传感器与微系统》2023年第5期143-146,151,共5页Transducer and Microsystem Technologies

基　　金：国家自然科学基金资助项目(617610380)。

摘　　要：针对单一室内定位技术的局限性,提出了一种基于扩展卡尔曼滤波(EKF)融合WiFi和行人航位推算(PDR)的定位方法。不同于传统WiFi指纹定位,本文基于随机森林(RF)模型建立多个基分类器,取投票结果的众数作为输出结果;通过采集手机内置传感器数据解算行人的步频、步长,并基于四元数进行航向估计。本文在EKF融合定位时,根据状态模型得到状态的预测值,RF模型输出观测值,根据观测值更新状态估计,推算下一时刻位置。试验表明,本文研究的融合算法的定位精度可达到1.26 m,比单一定位算法定位精度提高了1.07 m。Aiming at the limitation of single indoor positioning technology,a positioning method based on extended Kalman filtering(EKF)fuses WiFi and pedestrian dead reckoning(PDR)is proposed.Different from traditional WiFi fingerprint localization,multiple base classifiers based on random forest(RF)model is established,and the mode of voting results is taken as the output result.The step frequency and step length of pedestrians are calculated by collecting the data from the built-in sensor of mobile phone,and the heading is estimated based on quaternion.When EKF fusion positioning,the predicted value of the state is obtained according to the state model,the random forest model outputs the observed value,the state estimation is updated according to the observed value,and the position of the next time is calculated.Experimental result shows that the positioning precision of the fusion algorithm studied in this paper can reach 1.26 m,which is 1.07 m higher than that of a single positioning algorithm.

关 键 词：室内定位 WIFI 随机森林 行人航位推算 扩展卡尔曼滤波 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]