Indoor Tracking Persons Using Bluetooth: A Real Experiment with Different Fingerprinting-Based Algorithms

Abstract

In outdoor localization, global positioning systems (GPS) has been widely used. Indoor applications require a precise estimation that GPS can not achieve. Several technologies have been tried out as WI-FI, RFID, Bluetooth, Zigbee and others. This paper describes an experiment conducted in a medium-sized room in which six zones have been identified and two Bluetooth transmitters were installed. The aim is to enable continuous monitoring of areas where a person moves. For this purpose, we have used the technique of RSSI fingerprinting and tested three different algorithms. The best results were obtained with an algorithm based on SVM, which yielded success rates of 88.54%. Based on this algorithm, we intend to develop a cheap and easily configurable indoor localization system. © Springer International Publishing Switzerland 2013.