Semi-supervised Wi-Fi Flooring

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

Final Year Thesis Oral Defense

Title: "Semi-supervised Wi-Fi Flooring"

by

LI Mengxuan

Abstract:

Nowadays location is one of the most important context information in 
mobile and pervasive computing. Because of the recent proliferation of 
location-based services indoors, the need for an accurate floor prediction 
system which is easy to deploy in a mall-sized multi-floor building is 
more urgent than ever. Currently, the most widely used approach for 
localization is GPS. However, in a multi-floor building, it is extremely 
difficult to determine which floor the user is on using GPS signal. 
Other floor prediction systems that researchers have built include 
sensor-based systems and 3D-fingerprint system. Nevertheless, for these 
systems, the cost for deployment and maintenance are both very high, which 
significantly limits their large-scale use. In my final year thesis, I 
proposed a Wi-Fi flooring algorithm that only needs Wi-Fi signal. I used 
Wi-Fi signal, network embedding and modified agglomerative clustering 
algorithm to accomplish the Wi-Fi flooring task, and a floor prediction 
system used to estimate which floor the user is on inside a mall-sized 
multi-floor building can be developed based on my work. In comparison with 
the previous mentioned approaches, the floor prediction system based on my 
algorithm is cost-effective. Moreover, the ubiquitous Wi-Fi signal 
provides additional convenience for large-scale deployment of the systems 
based on my work and makes them more feasible.


Date            : 6 May 2021 (Thursday)

Time            : 15:00-15:40

Zoom Link:
https://hkust.zoom.us/j/97906871701?pwd=LzR6aGphdVBsak9iRGtrM25LREdkZz09

Meeting ID      : 979 0687 1701
Passcode        : 136636

Advisor         : Prof. CHAN Gary Shueng-Han

2nd Reader      : Prof. WONG Raymond Chi-Wing