基于RSS指纹的室内定位方法  

Indoor positioning method based on RSS fingerprint

作　　者：苏奎[1] 于曦[1] 芦思宇 SU Kui;YU Xi;LU Siyu(School of Medical Imaging,Mudanjiang Medical University,Mudanjiang 157011,Heilongjiang,China)

机构地区：[1]牡丹江医学院医学影像学院,黑龙江牡丹江157011

出　　处：《智能计算机与应用》2024年第10期158-163,共6页Intelligent Computer and Applications

基　　金：牡丹江市应用技术研究与开发计划项目(HT2022JG130)。

摘　　要：随着智能手机和WiFi设备变得日益普及,室内定位技术、特别是基于WiFi接收信号强度RSS的指纹定位方法在不断进步。K近邻算法是目前普遍应用在RSS指纹定位的快捷有效算法。这种方法因其成本效益低和易于部署的特性而备受青睐。然而室内环境复杂多变,多径效应、障碍物遮挡等问题常常导致RSS信号出现显著误差,影响定位精度。为提高其鲁棒性和准确性,在研究K近邻算法的同时通过滤波预处理、改进权重因子和加入冗余动态数据库等方式对原有方法进行改进。实验结果显示改进方法提高了楼层定位的准确性,并降低了平均定位误差。As smartphones and WiFi devices become increasingly popular,indoor positioning technology,especially the method based on WiFi Received Signal Strength fingerprints,is rapidly developing.The K-Nearest Neighbor algorithm is currently a widely used,fast,and effective algorithm for RSS fingerprint positioning.This method is favored for its cost-effectiveness and ease of deployment.However,the complex and variable indoor environment,along with issues such as multipath effects and obstructions,often lead to significant errors in RSS signals,affecting positioning accuracy.To improve its robustness and accuracy,the study improves upon the original method by filtering preprocessing,refining weight factors,and incorporating a redundant dynamic database alongside researching the K-Nearest Neighbor algorithm.Experimental results show that the improved method enhances the accuracy of floor positioning and reduces the average positioning error.

关 键 词：机器学习 室内定位 指纹匹配 RSS 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]