Off-campus UNL users: To download campus access dissertations, please use the following link to log into our proxy server with your NU ID and password. When you are done browsing please remember to return to this page and log out.

Non-UNL users: Please talk to your librarian about requesting this dissertation through interlibrary loan.

Misbehavior Detection and Privacy in Cellular Based Vehicular Communication Networks

Sohan Gyawali, University of Nebraska - Lincoln

Abstract

Vehicular communication networks are evolving rapidly with the emergence of novel technologies and are envisioned to support a variety of traffic safety and navigation applications. In this dissertation, misbehavior detection in cellular based vehicular communication networks is studied. In particular, challenges and solutions for security, privacy and misbehavior detection system (MDS) in cellular based vehicular communications are studied. A machine learning based MDS that is trained using realistic vehicular network datasets is presented and is compared with the previous methods in accurately identifying various misbehaviors in vehicular communication networks. Furthermore, machine learning and reputation based collaborative MDS is presented to ensure the reliability of both vehicles and messages and to further improve the detection accuracy. Reputation scores of the vehicles are taken as belief values and these belief values are combined using Dempster-Shafer (DS) theory. The results obtained from the DS theory and beta distribution is used to update the reputation score of vehicles. Moreover, dynamic reputation update policy in collaborative MDS is studied. In the proposed method, the DS theory is used by the local authority to predict the average number of true messages. A deep reinforcement learning policy is then used to determine the optimum reputation update policy to stimulate vehicles to send true feedbacks. In addition to collaborative MDS and dynamic reputation update policy, privacy-preserving MDS is proposed. The proposed privacy-preserving MDS can identify misbehavior without violating the privacy of the vehicle. Encrypted feedbacks from vehicles are aggregated at the local authority using additive homomorphic properties and without violating the privacy of the vehicle. The decryption of aggregate feedback is done at the TA to securely update the reputation score of the vehicle. Finally, open issues and possible future research directions in the misbehavior detection and privacy of cellular based vehicular communication networks are highlighted in this dissertation.

Subject Area

Computer Engineering

Recommended Citation

Gyawali, Sohan, "Misbehavior Detection and Privacy in Cellular Based Vehicular Communication Networks" (2020). ETD collection for University of Nebraska-Lincoln. AAI28027807.
https://digitalcommons.unl.edu/dissertations/AAI28027807

Share

COinS