Performance Improvements in Software-Defined Vehicular Ad Hoc Networks

Abstract

In this dissertation, we investigate performance improvements in software-defined and virtualized vehicular ad-hoc networks (VANETs) with advanced technologies. We firstly present a deep reinforcement learning (DRL) approach in software-defined vehicular ad-hoc networks with trust management. In this research, we propose to use a deep Q-learning (DQL) approach and the centralized control mechanism of SDN to address the bad influences of the malicious nodes in VANETs. The trust of each vehicle and the reverse delivery ratio are considered in a joint optimization problem, which is modeled as a Markov decision process with state space, action space, and reward function. DQL is used to obtain the best link quality policy under the bad influence of the malicious vehicles in the inter-vehicle communication for ITS. Secondly, we focus on distributed SDN and blockchain technology for connected vehicles in smart cities. Specifically, we propose a novel blockchain-based hierarchical distributed software-defined VANET framework (block-SDV) to establish a secure and reliable architecture that operates a distributed way to overcome the security issues of VANETs. Thirdly, we propose a novel framework of blockchain-based mobile edge computing for a future VANET ecosystem (BMEC-FV), in which we have adopted a hierarchical architecture. In the underlying VANET environment, we propose a trust model to ensure the security of the communication link between vehicles.