Blockchain-Based Distributed Software-defined Vehicular Networks: A Dueling Deep Q-Learning Approach

Vehicular ad hoc networks (VANETs) have become an essential part in smart transportation systems of modern cities. However, because of dynamicity and infrastructure-less of VANETs, the ever increasing number of network security issues become obstacles for the realization of smart cities. Software-defined VANETs have provided a reliable way to manage VANETs dynamically and securely. However, the traditionally centralized control plane makes it vulnerable to malicious nodes and results in performance degradation. Therefore, a distributed control plane is necessary. How to reach a consensus among multiple controllers under complex vehicular environment is an essential problem. In this paper, we propose a novel blockchain-based distributed software-defined VANET framework (block-SDV) to establish a secure architecture to overcome the above issues. The trust features of blockchain nodes, the number of consensus nodes, trust features of each vehicle, and the computational capability of the blockchain are considered in a joint optimization problem, which is modeled as a Markov decision process with state space, action space and reward function. Since it is difficult to be solved by traditional methods, we propose a novel dueling deep Q-learning (DDQL) with prioritized experience replay approach. Simulation results are presented to show the effectiveness of the proposed block-SDV framework.