11 / 58
10
HIND OBAID HAMAD SAEED
AL FALASI
Department of Computer and Network Engineering
College of Information Technology
Title
Using Similarity to Achieve Trust to Enhance Decision Making in Vehicular Safety Applications
Faculty Advisor
Dr. Hesham El-Sayed
Defense Date
24 November 2015
Abstract
Vehicles exchange different types of messages either periodically or as needed for different types of
applications. The data in the network of vehicles can be used to extract valuable knowledge to support
various applications in vehicular ad hoc networks (VANETs). Knowledge gained from the gathered data
can be used to create local views of the network for individual vehicles; for instance, a vehicle can form a
view of a subset of the network using neighboring vehicles’ directions of travel, speeds, and the types of
applications they run. In the network, vehicles that have common attributes and requirements facilitate
the establishment of trust between them as these shared features make up the foundation for trust. Trust
relations between vehicles can be utilized for enhancing the performance and reliability of some applications.
This dissertation is concerned with trust establishment in VANETs, and how it can be utilized to enhance
efficiency and decision making in the network. We provide solutions to this question: How to utilize trust
relationship between vehicles to improve decision-making and efficiency in VANET safety applications.
In our research, we aim to establish trust relationships through similarity to assist vehicles in identifying
false safety messages in the network. We start by designing and implementing a trust management
system to generate and process trust values and to establish a set of trusted relationships for vehicles
running vehicular safety applications. Next, we explore the possibility of enhancing the decision-making
process using trust. First, we develop an analytical model that associates trust with the performance of
the decision-making process and the accuracy, and then we study the effectiveness of similarity-based
trust in identifying false safety event messages in VANETs. Finally, we show that similarity-based trust has
a positive impact on the time needed to make a decision and on the accuracy of that decision.
Dissertation