Electronic Thesis and Dissertation Repository

Degree

Master of Engineering Science

Program

Electrical and Computer Engineering

Supervisor

Dr. Xianbin Wang

Abstract

The dramatic increase in the number of connected devices and the significant growth of the network traffic data have led to many security vulnerabilities and cyber-attacks. Hence, developing new methods to secure the network infrastructure and protect data from malicious and unauthorized access becomes a vital aspect of communication network design. Intrusion Detection Systems (IDSs), as common widely used security techniques, are critical to detect network attacks and unauthorized network access and thus minimize further cyber-attack damages. However, there are a number of weaknesses that need to be addressed to make reliable IDS for real-world applications. One of the fundamental challenges is the large number of redundant and non-relevant data. Feature selection emerges as a necessary step in efficient IDS design to overcome high dimensionality problem and enhance the performance of IDS through the reduction of its complexity and the acceleration of the detection process. Moreover, detection algorithm has significant impact on the performance of IDS. Machine learning techniques are widely used in such systems which is studied in details in this dissertation. One of the most destructive activities in wireless networks such as MANET is packet dropping. The existence of the intrusive attackers in the network is not the only cause of packet loss. In fact, packet drop can occur because of faulty network. Hence, in order detect the packet dropping caused by a malicious activity of an attacker, information from various layers of the protocol is needed to detect malicious packet loss effectively. To this end, a novel cross-layer design for malicious packet loss detection in MANET is proposed using features from physical layer, network layer and MAC layer to make a better detection decision. Trust-based mechanism is adopted in this design and a packet loss free routing algorithm is presented accordingly.

mahsa.pdf (15 kB)

Available for download on Saturday, September 01, 2018

Share

COinS