Improving the Routing Process in Vehicular Networks While Considering Both Security and Quality of Service (QoS) through the Use of Machine Learning and Fuzzy Logic, within the Framework of Fog Computing Architectures and Software-Defined Networking (S

Abstract

There are several challenges for routing protocols, some of which are associated with security concerns and others are with Quality of Service (QoS). The highly dynamic and distributed nature of the network topology further exacerbates these challenges. This paper aims to offer a comprehensive solution that concurrently tackles security and QoS issues. After an authentication phase, machine learning classifiers installed in the SDN controller and fog nodes are used to analyze network activity behaviorally. Under a load-balancing policy controlled by the software-defined controller, these nodes process data based on priority levels. An adaptive aerial path is incorporated into a reinforcement learning mechanism for route discovery that is conditioned by security and QoS standards. The suitability of the relay drone is assessed using an applied fuzzy logic technique. Using computer simulations, the suggested model was assessed according to several important performance factors, such as attack success rate, packet delivery ratio, throughput, and end-to-end delay. These factors have been tested under varying parameters such as the number of vehicles in the network, the proportion of malicious nodes, and vehicle speed. The outcomes showed that the suggested model outperformed the benchmark protocol (QL-TRT).