Efficient Locality Classification for Indoor Fingerprint-based Systems

MPhil Thesis Defence


Title: "Efficient Locality Classification for Indoor Fingerprint-based Systems"

By

Mr. Ka Ho CHOW


Abstract

Locality classification is an important component to enable location-based 
services. It entails two sequential queries: 1) whether a target is within the 
site or not, i.e., inside/outside region decision, and 2) if so, which area in 
the region the target is located, i.e., area classification. Locality 
classification is hence more coarse-grained and efficient as compared with 
pinpointing the exact target location in the region. The classification problem 
is challenging, because fingerprints may not exist outside the region for 
training. Furthermore, the target may sample an incomplete RSSI vector due to, 
say, random signal noise, momentary occlusion or scanning duration. The 
algorithm also has to be computationally efficient. We propose INOA, a scalable 
and practical locality classification overcoming the above challenges. INOA may 
serve as a plug-in before any fingerprint-based localization, and can be 
incrementally extended to cover new areas or regions for large-scale 
deployment. Its preprocessor cherry-picks only those discriminating access 
points, which greatly enhances computational efficiency and accuracy. By 
formulating a “one-class” classifier using ensemble learning, INOA accurately 
decides whether the target is within the region or not. Extensive experimental 
trials in different sites validate the high efficiency and accuracy of INOA, 
without the need of full RSSI vectors collected at the target.


Date:			Monday, 25 June 2018

Time:			2:30pm - 4:30pm

Venue:			Room 5560
 			Lifts 27/28

Committee Members:	Prof. Gary Chan (Supervisor)
 			Prof. Dik-Lun Lee (Chairperson)
 			Dr. Raymond Wong


**** ALL are Welcome ****