Evaluation of Indoor Positioning System with Bluetooth Low Energy Technology

Choong, Melcom Pak Hin (2025) Evaluation of Indoor Positioning System with Bluetooth Low Energy Technology. Final Year Project (Bachelor), Tunku Abdul Rahman University of Management and Technology.

Text Melcom Choong Pak Hin -FULL TEXT.pdf Restricted to Registered users only Download (7MB)

Abstract

This project explores the development of a Bluetooth Low Energy (BLE)-based Indoor Positioning System (IPS) aimed at overcoming the limitations of GPS in enclosed environments. The system estimates location by utilising Received Signal Strength Indicator (RSSI) values transmitted from BLE beacons and received by multiple BLE gateways. A fingerprinting technique is used to map RSSI values to specific indoor locations, with a K-Nearest Neighbour (KNN) algorithm for location prediction and a Moving Average (MA) filter to reduce RSSI signal fluctuations. The system was implemented and tested in a classroom environment using 32 reference points. Three configurations were evaluated: (1) Euclidean Distance Only, (2) Euclidean Distance with Moving Average Filter, and (3) Euclidean Distance with Moving Average Filter and KNN. The best result was achieved using the third configuration, which obtained an average estimation accuracy 70.156%, demonstrating high accuracy well within the 2-meter target range. The findings confirm that integrating fingerprinting, KNN, and MA filtering significantly improves the reliability and accuracy of BLE-based indoor positioning. This system is suitable for practical applications such as asset tracking, attendance monitoring, and crowd management in GPS-denied environments. Overall, the project contributes to the development of accurate, low-cost, and energy-efficient indoor positioning solutions.