Towards Ubiquitous Indoor Localization Service via Multi-Modal Sensing on Smartphones

The Hong Kong University of Science and Technology
Department of Computer Science and Engineering


PhD Thesis Defence

Title: "Towards Ubiquitous Indoor Localization Service via Multi-Modal Sensing 
on Smartphones"

By

Mr. Han XU


Abstract

Indoor localization is of great importance to a wide range of applications in 
this era of mobile computing, attracting extensive research effort over recent 
decades. Current mainstream solutions rely on Received Signal Strength (RSS) of 
wireless signals as fingerprints to distinguish and infer locations. However, 
those methods suffer from fingerprint ambiguity that roots in multipath fading 
and temporal dynamics of wireless signals, which invalidate theoretical 
propagation models, distort received signal signatures, and fundamentally 
constrain the performance of indoor localization. With the trend moving towards 
equipment of smart devices in daily life and adoption of enhanced sensors,  we 
identify the opportunity of ubiquitous indoor localization service via the 
multi-modal sensing abilities on smartphones. Firstly, we propose Argus, an 
image-assisted localization solution for mobile devices by harnessing their 
Visual Sensing abilities. The basic idea of Argus is to extract geometric 
constraints from crowdsourced photos, and to reduce fingerprint ambiguity by 
mapping the constraints jointly against the fingerprint space. Secondly, we 
design TUM, an Acoustic Sensing localization scheme Towards Ubiquitous 
Multi-device localization. The basic idea of TUM is to utilize the 
dual-microphones and speakers to obtain distance cues among devices, while 
resolving the localization ambiguity with the help of MEMS sensors. Thirdly, we 
exploit the Inertial Sensing abilities on smartphones and propose RAD. The 
basic idea is to automatically generate a fingerprint database through space 
partition, while achieving fine-grained localization via a discretized particle 
filter with sensor data fusion. Finally, we design an indoor localization 
system ClickLoc that achieves sub-meter accuracy by harnessing Multi-Modal 
Sensing abilities on smartphones. We prototype the above schemes with commodity 
devices, and evaluate their performances in various indoor environments. 
Experimental results demonstrate improved indoor localization accuracy, better 
user interaction and less overhead compared with classical RSS-based schemes.


Date:			Wednesday, 19 October 2016

Time:			4:00pm - 6:00pm

Venue:			Room 1504
 			Lifts 25/26

Chairman:		Prof. Jiheng Zhang (IELM)

Committee Members:	Prof. Ke Yi (Supervisor)
 			Prof. Gary Chan
 			Prof. Lei Chen
 			Prof. Xiangtong Qi (IELM)
 			Prof. Jianping Wang (Comp. Sci., CityU)


**** ALL are Welcome ****