Year: 2008
Type: dissertation
Abstract: Security and trust are two interdependent concepts and are often used \ninterchangeably when defining a secure wireless sensor network (WSN) system. \nHowever, security is different from trust in that, it assumes no node is trustworthy \nand requires ongoing authentication using sophisticated protocols leading to high \ncommunication and computation overheads. This makes the traditional \ncryptographic security tools hard, if not impossible, to be used in wireless sensor \nnetworks that are severely resource constrained. Trust on the other hand is the \nexact opposite of security in that any node can interact with any other and requires \nno authentication and unwrapping of hidden keys to carry on with their business \nand hence carries zero overhead. However, this leads to the miss-use and abuse of \nnetworks causing loss and damage to the owners of the networks. This thesis \nfocuses on developing novel methods for modelling and managing trust that \nenable WSN to be secure while significantly reducing computing and \ncommunication overheads. \nAlthough researchers have been studying the problem of trust modelling and \nmanagement in wireless sensor networks for over a decade, their focus was on the \ntrust associated with routing messages between nodes (communication trust). \nHowever, wireless sensor networks are mainly deployed to sense the world and \nreport data, both continuous and discrete. However, there are no methods in the \nliterature that focus on the trust associated with misreporting data (data trust). In \nthis thesis, we model the trust associated with the integrity of the data, and \npropose methods to combine the data trust with the communication trust to infer \nthe total trust. \nBayesian probabilistic approach is used to model and manage trust. A new risk \nassessment algorithm for establishing trust in wireless sensor networks based on \nthe quality of services characteristics of sensor nodes, using the traditional \nweighting approach is introduced. Then a Beta distribution is used to model \ncommunication trust (due to its binary nature) and determine the weights in terms \nof the Beta distribution parameters to probabilistically combine direct and indirect \ntrust. \nThe thesis extends the Bayesian probabilistic approach to model data trust for \ncases when the sensed data is continuous. It introduces the Gaussian trust and \nreputation system to that accounts for uncertain characteristics of sensor data. \nFinally we introduce a Bayesian fusion algorithm to combine the data trust and \ncommunication trust to infer the overall trust between nodes. Simulation results \nare presented to demonstrate how the models accurately classify different nodes as \nbeing trustworthy or not based on their reliability in sensor reporting and routing \nfunctions.
Author: MM Momani
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2405484291