University of Twente Student Theses


Practical indoor localization using bluetooth

Scheerens, Daan (2012) Practical indoor localization using bluetooth.

[img] PDF
Abstract:Localization is a problem that has been addressed using a variety of technologies. In this thesis the use of Bluetooth for indoor localization is studied. The advantage of this technology over others is that it is pervasively available, is relatively cheap and has a relatively low power consumption. Especially the fact that Bluetooth is integrated in a wide range of mobile devices, makes its use attractive. The question being answered by this thesis is which design of a Bluetooth based localization system works well for indoor environments. The context of the localization system is that of an o�ce building in which the location of employees is tracked. The main contribution of this thesis is a practical evaluation of Bluetooth as a technology for indoor localization. Received Signal Strength in the inquiry phase of the Bluetooth device discovery protocol has been identified as the most suitable localization measure. This measure, however, has the disadvantage that the sample rate is relatively low. Tests have shown that at least one minute is required to collect a sufficient number of samples. Because of the relatively low sample rate, accurate location estimation of moving people is not possible. To study the localization performance of Bluetooth for indoor environments a number of localization algorithms were tested. These algorithms include: Ecolocation, calibrated and uncalibrated Log-Normal Shadowing model based algorithms and �ngerprinting based algorithms. For each algorithm localization accuracy was computed using datasets which were collected in a test environment. Furthermore, the e�ect of several controllable and uncontrollable parameters on localization accuracy was tested for the algorithms. The controllable parameters that were tested are: number of access points and window size, and the uncontrollable parameters that were tested are: device orientation, device height, transmitter power level and environment structure. An analysis of the e�ects of these parameters shows that the uncalibrated algorithms are less sensitive to the uncontrollable parameters. These algorithms, however, require more access points to achieve reasonable localization accuracy. The main conclusion of this thesis is that the uncalibrated localization algorithms are best suitable for indoor localization. This is because of the low impact of the uncontrollable parameters on their localization performance. Which algorithm works the best depends on the number of access points that can detect a target. For 5 access points or less, Ecolocation appears to yield the best location estimates. Otherwise, the uncalibrated Log-Normal Shadowing model based algorithm performs the best.
Item Type:Essay (Master)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science MSc (60300)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page