An Intelligent System Using Deep Learning-Based Link Quality Prediction and Optimization Enabled Secure Communication in UAV Network

Abstract

Recently, unmanned aerial vehicles (UAVs) are widely deployed, due to their emerging benefits, like less cost, and on-demand installation. The exchanging of information is done amongst the UAVs. However, this kind of communication may contain security issues because of its vibrant network topology. For addressing these problems, this article devises an optimized model for attaining secure communication amongst UAVs. The quality of the link is predicted using Deep Q Network using a prediction agent. The establishment of effective routes is done with Tunicate Honey Badger Algorithm. The fitness is modeled with link quality and distance. The evaluation and monitoring agent is employed for data communication. The maintenance agent helps to manage the route failures. The decision-making agent utilizes Shepard Convolutional Neural Network (ShCNN) for detecting the malicious behaviors. The ShCNN employs certain parameters as input to detect attack. If attack is encountered, then defensive agent alleviates identified attack.