CONTJUNCTION: INDOOR LOCALIZATION BASED ON JUNCTION OF SIGNAL CONTOURS

MPhil Thesis Defence


Title: "CONTJUNCTION: INDOOR LOCALIZATION BASED ON JUNCTION OF SIGNAL 
CONTOURS"

By

Mr. Weipeng ZHUO


Abstract

Due to its ease of deployment, WiFi-based fingerprinting has emerged as a 
promising technique for indoor localization. Traditional fingerprinting 
algorithms suffers that due to signal noise, disperse reference points may 
be picked for location estimation, resulting in large location error. In 
this paper, we propose a localization scheme termed "Contjunction", which 
avoids the problem by leveraging WiFi signal contour intersections. We try 
to identify a reference region where the most of signal contours 
representing the target RSSI intersect. After the region is found, 
target's position is estimated based on k-nearest-neighbor approach. A 
generic distance metric is proposed to measure the distance between 
reference point and target in the signal space. Closed-form examples are 
given for Uniform, Normal and Log-normal distributions. Simulation results 
show that our scheme outperforms others from recent literature by more 
than 30%. Experimental results from the Hong Kong International Airport 
(HKIA) and HKUST campus show that our proposed scheme achieves at least 
30% improvement. In addition, an APP is built on Android platform with our 
algorithm implemented, which we have used for several demos in the 
airport.


Date:			Monday, 12 August 2013

Time:			4:00pm – 6:00pm

Venue:			Room 5508
 			Lifts 25/26

Committee Members:	Prof. Gary Chan (Supervisor)
 			Prof. James Kwok (Chairperson)
 			Dr. Pan Hui


**** ALL are Welcome ****