Indoor positioning using semi supervised fingerprint building technique based on crowd sourced information

• Main Author: Toh, Cornelius Dong Tou
• Format: Thesis
• Published: 2020
• Subjects: QA76 Computer software; TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://eprints.sunway.edu.my/2398/

Indoor positioning using semi-supervised fingerprint building technique based on crowdsourced information, where costly and extensive site survey in supervised fingerprint building technique will be reduced while able to detect if a person is at a specified location when a person has longer stay and having better initial location recognition than unsupervised fingerprint building. The system is also able to label locations automatically using unsupervised fingerprint building technique with the aid of supervised fingerprint building technique by recognising first locations thru fingerprints built thru data from site survey, and then labelling the upcoming neighbouring locations thru the neighbour list. With the help of users’ information based on their device’s models, where they will be able to collect network data of the location they are currently at and contributes WiFi Access Point data to create useful information for location recognition based on their device models to improve accuracy consistency across different models. This hybrid fingerprint building technique is aimed to create a system that has the best of both supervised and unsupervised fingerprint building technique to eliminate the disadvantage of each technique with their respective advantages. The research is carried out in Open Area, such as the Sunway Pyramid Mall where locations are close to each other in a vast area, and Closed Area, Sunway University which has walls and rooms to better isolate wireless signals. The results showed improvements over supervised and unsupervised fingerprint building techniques.