基于LoRa指纹的室内定位算法

在基于LoRa的室内定位研究中,提出一种基于LoRa指纹和支持向量回归(SVR)的室内定位算法。针对传统基于无线信号RSSI指纹和SVR室内定位算法定位精度不高问题,从两个方面进行改进:在指纹特征方面,增加LoRa测距指纹,提高指纹稳定性;在指纹数据库建立和在线定位过程中,分别采用高斯滤波和中位数滤波来对指纹进行预处理,消除指纹的粗大误差。实验结果显示:1 m以内的定位误差的累积概率为78.5%,3 m以内的定位误差的累积概率为90%。增加LoRa测距指纹之后定位精度相比之前提高了40%;增加了高斯滤波与中位数滤波预处理后定位精度较传统的支持向量回归算法提高了38%。两个方面改进之后定位精度总体提高63%,证明了该算法的两个改进是有效的。

INDOOR LOCATION ALGORITHM BASED ON LORA FINGERPRINT

This paper studied indoor location based on LoRa,and proposed an indoor location algorithm based on LoRa fingerprint and support vector regression(SVR).Aiming at the low location accuracy of traditional RSSI fingerprint and SVR indoor location algorithm based on wireless signals,two improvements were made in this paper.In the aspect of fingerprint characteristics,LoRa ranging fingerprint was added to improve fingerprint stability;in the process of establishing fingerprint database and online locating,Gauss filter and median filter were used to pre-process fingerprints to eliminate the gross errors of fingerprints.The experimental results show that the cumulative probability of location error within 1 m is 78.5%,and the cumulative probability of location error within 3 m is 90%.After adding LoRa ranging fingerprint,the location accuracy is improved by 40%compared with before;after adding Gauss filter and median filter,the location accuracy is improved by 38%compared with traditional support vector regression algorithm.After two improvements,the location accuracy is improved by 63%,which proves that the two improvements of the algorithm are effective.