Outdoor Fingerprinting Localization Using Sigfox

Abstract

The Internet of Things (IoT) has caused the modern society to connect everything in our environment to a network. In a myriad of IoT applications, smart devices need to be located. This can easily be done by satellite based receivers. However, there are more energy-efficient localization technologies, especially in Low Power Wide Area Networks (LPWAN). In this research, we discuss the accuracy of an outdoor fingerprinting technique using a large outdoor Sigfox dataset which is openly available. A kNN (k Nearest Neighbors) algorithm is applied to our fingerprinting database. 31 different distance functions and four RSS data representations are evaluated. Our analysis shows that a Sigfox transmitter can be located with a mean estimation error of 340 meters.