The School of Computer Science at the University of Windsor is pleased to present …
Intrusion Detection System for Basic Safety Messages in Connected Vehicles Using Transfer Learning Approach
PhD. Seminar by: Muhammad Anwar Shahid
Date: Tuesday, August 15, 2023
Time: 9:30 AM -11:30AM
Location: Essex Hall, Room #122
Abstract:
Intelligent Transportation System (ITS) is a promising technology to make safer, more coordinated and smarter use of transportation networks. The main objective of ITS is to ensure better traffic efficiency, fuel conservation and reducing road accidents through Vehicular Ad Hoc Network (VANET). VANET is a special type of Mobile Ad Hoc Network (MANET), which consists of cars as network nodes, Road Side Unit (RSU) and On Board Unit (OBU). The US Department of Transportation has estimated that vehicular communication based on Dedicated Short Range Communication (DSRC) may reduce traffic crashes by up to 82%, leading to a significant reduction in loss of life as well as economic losses. Vehicles in VANET share sensitive information with other nearby vehicles or infrastructure through Basic Safety Messages (BSM). A BSM contains critical information such as vehicle identity, position, speed, heading, and other parameters. With the sheer volume of vehicles on the road and the rising popularity of connected vehicles, it is essential to secure communication in VANET against the various types of attackers. VANETs are exposed to various known threats and attacks such as bogus information, DoS, Sybil, Data Replay and many more. Many researchers around the globe are working to develop Intrusion Detection Systems (IDS) and strategies to secure communications in VANET. In our work, we propose transfer learning-based IDS for vehicle-to-vehicle (V2V) communication in VANET, which can not only detect malicious BSMs/senders but also, reduce the training time. We performed our experiment on the VeReMi extension dataset, a well-known dataset for V2V communication, using the structure and unstructured instances. Our framework showed better performance in terms of accuracy, precision, recall and F1-score as compared to other approaches.
Keywords: ITS, VANET, BSM, Transfer Learning, IDS, Deep Learning
PhD Doctoral Committee:
Internal Reader: Dr. Boubakeur Boufama
Internal Reader: Dr. Imran Ahmad
External Reader: Dr. Ning Zhang
Advisor(s): Dr. Arunita Jaekel