Federated Learning Framework for Privacy Security and Trust in Vehicular Networks
Loading...
Date
item.page.authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The emergence of the fifth-generation communications networks (5G) has brought a
newlinetechnological revolution to the world. It has laid a solid foundation for the vigorous
newlinedevelopment of the vehicular ad hoc networks (VANETs). VANET is self-managed network
newlinefor vehicles whose main motive is to improve the safety of driver and passenger.
newlineIntercommunication within the network helps to pass safety and traffic related information
newlineamong the vehicles. However, while communication performance has enhanced, Security,
newlineprivacy and Trust issues have increased. Attacks have become more aggressive, and
newlineattackers have become more strategic, which can lead to hazardous situations for drivers and
newlinepassengers. Current cybersecurity infrastructures are failing to keep pace with the
newlinesophistication of modern cyber-attacks, such as position falsification, spoofing, denial-of
newlineservice (DoS), distributed denial-of-service (DDoS) and internal attacks. However, detecting
newlinethese attacks and attackers is challenging and requires designing sophisticated and intelligent
newlinedetection system. There is a critical need for a security paradigm that not only anticipates
newlineand mitigates known threats but also evolves adaptively to counteract emerging
newlinevulnerabilities while maintaining user privacy, security and trust along with system
newlineperformance. This thesis presents a novel approach of introducing the federated learning to
newlinepreserve and secure the trusted communication among the vehicles and protect from the
newlineattacks and attackers. Federated learning is an emerging privacy preservation machine
newlinelearning paradigm proposed to enhance the privacy and security of the vehicular Networks.
newlineIt enables multiple vehicles to collaboratively train a global model without sharing their local training data, thus preventing individuals data from being revealed.
newline