摘要
针对传统Wi-Fi室内定位系统存在的数据冗余度高和楼层识别精度低两个问题,该文提出了一种基于双频Wi-Fi信息和AP选择的三维室内定位系统FI_EWKNN,定位系统包含楼层识别和平面定位两个模块。楼层识别模块通过分析双频RSSI分布特征构建置信区间指纹库,计算接收RSSI在各个楼层的置信度便可确定目标楼层,确定楼层后根据AP互信息筛选目标楼层质量较好的AP,最后利用EWKNN获得最终定位结果。实验结果表明,在大型复杂的室内场景下,提出的基于双频Wi-Fi信息的楼层识别准确率显著优于现有的单频Wi-Fi方法,识别准确率为98.6%。同时,与KNN,WKNN,E-WKNN算法相比,提出的定位算法在平均定位精度上分别提高了0.946m、0.845m和0.574m,有效提高了定位精度。
Aiming at the problem that the high data redundancy and low floor recognition accuracy issues in traditional Wi-Fi indoor positioning systems,a three-dimensional indoor positioning system named FI_EWKNN is proposed.The algorithm is based on dual-band Wi-Fi information and AP selection.The system consists of two modules:floor recognition and planar positioning.The floor identification module constructs a confidence interval fingerprint library by analyzing the distribution characteristics of dual-frequency RSSI,calculates the confidence degree of the received RSSI at each floor to determine the target floor,and then selects the AP with better quality at the target floor according to the AP mutual information after determining the floor.Finally,the final positioning result is obtained using EWKNN.Experimental results demonstrate that,in large and complex indoor environments,the proposed three-dimensional indoor positioning algorithm based on dual-band Wi-Fi information outperforms single-band Wi-Fi positioning methods in terms of floor recognition accuracy,achieving a recognition accuracy of 98.6%.Furthermore,compared to KNN,WKNN,and E-WKNN algorithms,the proposed solution achieves an improvement of 0.946m,0.845m,and 0.574m,respectively,in average positioning accuracy,effectively enhancing the accuracy of positioning.
作者
段肖铭
余科根
林贻若
DUAN Xiaoming;YU Kegen;LIN Yiruo(School of Environment Science Informatics,China University of Mining and Technology,Xuzhou,Jiangsu 221116,China)
出处
《测绘科学》
CSCD
北大核心
2023年第12期20-28,共9页
Science of Surveying and Mapping
关键词
室内定位
三维定位
双频Wi-Fi
楼层识别
AP选择
indoor positioning
three-dimensional positioning
dual-band Wi-Fi
floor identification
AP selection
作者简介
段肖铭(1999—),男,河南漯河人,硕士研究生,主要研究方向为室内定位与导航。E-mail:TS21160008A31@cumt.edu.cn;通信作者:余科根,教授,E-mail:kegen.yu@foxmail.com
