Wireless Sensor Network Fault Localization

MPhil Thesis Defence


Title: "Wireless Sensor Network Fault Localization"

By

Mr. Mian Qin


Abstract

Wireless sensor network (WSN) is a widely used distributed system to sense and 
interact with the physical world. Like other applications, wireless sensor 
network applications can be plagued by a variety of software faults. These 
faults can be critical and costly in debugging and deployment of sensor 
networks. Furthermore, these WSN software faults are difficult to locate. 
Therefore, automatic fault localization before deployment can significantly 
reduce the cost of debugging and deployment of a sensor network system. In 
general, debugging is composed of five steps: finding failure-triggering test 
cases, simplification of failed test cases, fault localization, repairing and 
regression testing. In another dimension, faults can reside in application 
logic, driver of on-board device and networking. Among these dimensions, fault 
localization has been considered the most important step of debugging, and 
application logic is more complex and bug-prone. We therefore focus on the 
study of fault localization of the application logic faults. In this thesis, 
first, an overview of the thesis will be presented. Then, WSN applications and 
automatic fault localization techniques will be introduced. After that, based 
on the state-of-art fault localization technique Tarantula, an adapted 
algorithm has been proposed in order to increase its effectiveness. We also 
evaluated this algorithm by conducting controlled experiments. As a result, the 
effectiveness of Tarantula has been improved. Based on this algorithm, we have 
also built an infrastructure for before-deployment fault localization on sensor 
network programs. This infrastructure provides various functions for different 
WSN debugging tasks performed by both developers and researchers.


Date:				Tuesday, 23 December 2008

Time:				2:00p.m.-4:00p.m.

Venue:				Room 3501
 				Lifts 25-26

Committee Members:		Dr. Shing-Chi Cheung (Supervisor)
 				Prof. Lionel Ni (Chairperson)
 				Prof. Vincent Shen



**** ALL are Welcome ****