IoT Architecture Design: Network, Security and Application

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


PhD Thesis Defence


Title: "IoT Architecture Design: Network, Security and Application"

By

Mr. Zeyu WANG


Abstract

The Internet of Things (IoT) has quickly moved from hype to reality. Like other 
disruptive technologies, IoT holds the potential for societal scale impact by 
transforming many industries as well as our daily lives. IoT would benefit the 
world significantly if the generated data from them can be shared securely and 
efficiently.

In this thesis, we mainly focus on the following topics:

1. Lightweight indoor localization. Indoor localization is a promising 
application in IoT scenario. Current visible light positioning system suffers 
the problem of overhead. We first propose a visible light positioning system 
using polarization modulation. Our system can work with non-intelligent light 
other than general LED and achieves high localization accuracy with low-end 
smartphones. We also design another visible light positioning system with an 
ambient light sensor which leverages the property of frequency aliasing and 
dynamic frequency adjustment. Since the power of the ambient light sensor is 
very low, it can enable indoor location inside a smartwatch or other low power 
wearables.

2. Secure authentication of IoT devices. We discover a vulnerability in one 
popular home automation brand related to the authentication procedure of IoT 
devices. Based on this study, we propose a proximity-based mechanism for 
smartphones to authenticate IoT devices which combines large RSS-variation 
detection and matching between RSS-trace and smartphone's sensor-trace, to 
perform reliable proximity detection

3. Scalability, security, and performance of blockchain. The blockchain is a 
promising approach to make IoT data more valuable. However, current blockchain 
schemes are lack of scalability and face the problem of security. To handle 
this problem, we propose a reputation-based secure, high incentive and fast 
blockchain system to handle the problem of scalability and security. In this 
work, we leverage a Raft-based synchronous consensus protocol and a 
double-chain architecture to guarantee its performance and security. After 
that, we also investigate the selfish behavior in the blockchain system which 
may limit the system performance. We propose a layer-2 reputation protocol that 
can be deployed on nowadays blockchain system to solve the problem of 
selfishness. In this work, we design a two-steps gossip protocol with a robust 
reputation calculation scheme to achieve global consensus on every node's 
reputation score efficiently and securely.


Date:			Friday, 18 January 2019

Time:			2:00pm - 4:00pm

Venue:			Room 3494
 			Lifts 25/26

Chairman:		Prof. Zhengtang Luo (CBE)

Committee Members:	Prof. Qian Zhang (Supervisor)
 			Prof. Kai Chen
 			Prof. Lei Chen
 			Prof. Wei Zhang (ECE)
 			Prof. Minghua Chen (CUHK)


**** ALL are Welcome ****