Institutional Repository of Vilnius University
Bluetooth paremtos navigacijos sistemos algoritmų tyrimas /
Abstract
dc:descriptionThere are two main methods used to determine your exact position using a Bluetooth connection: triangulation and fingerprinting. This work will delve into fingerprinting approach by applying different classification algorithms. This work will delve deeper into fingerptinting approach by applying different classification algorithms and try to separate the most optimal and accurate algorithm. Several algorithms are discussed and tested: support vector machine, k-nearest neighbor algorithm, weighted k-nearest neighbor algorithm. Also, an attempt was made to develop an algorithm using a weighted k-nearest neighbor algorithm and taking into account human motion in space. The accuracy of the algorithms was assessed using standard and alternative assessment metrics. Alternative assessment metrics take into account the inaccuracy of the generated decision, in other words how many sectors are allocated to the correct answer. All the above algorithms were evaluated in both ways and the weighted k-nearest neighbor algorithm had the highest accuracy. The second most accurate was the standard k-nearest neighbor algorithm, followed by support vector machine algorithm. The developed algorithm that takes into account human movement in space has not payed off. The resulting accuracy rate is lower than commonly applied algorithms. This may be due to a small data set and a lack of additional mobile phone sensor readings.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2021
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Garnevičius, Aurimas,
- Contributors dc:contributor
-
- Raila, Tomas
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD88378752&prefLang=en_US