Stochastic Packet Forwarding Algorithm in Flying Ad Hoc Networks

Abstract

The drones, also officially referred to as unmanned airborne vehicles (UAVs), have captured the attention of hobbyists, researchers, and investors, and are becoming increasingly popular for various commercial, industrial, and public-safety applications. As an essential ingredient of Internet-of-Drones, Flying Ad Hoc Networks (FANETs) largely consisting of various drones are expeditiously proliferating and playing an important role in realizing the goal of coordinating the access of drones to controlled airspace and providing navigation services. However, packet forwarding in FANETs is challenged by the unique characteristics of FANETs, such as unstable wireless medium and intermittent connectivity caused by high mobility of drones. In this paper, we propose a stochastic packet forwarding algorithm, also called SPA, to provide efficient and reliable data transmission in FANETs. The basic idea of the SPA is to make a stochastic forwarding drone selection based on the combination of multiple real-time network metrics. By objectively allocating the weight to multiple real-time network metrics based on the entropy weight theory, the SPA computes the forwarding availability of each forwarding candidate drone. Then, the forwarding probability of each forwarding candidate drone is calculated, and the forwarding drone is stochastically chosen from all forwarding candidate drones based on the calculated forwarding probability. In experimental performance evaluation, we select link throughput and link expiration time as real-time network metrics, and evaluate the proposed stochastic packet forwarding algorithm through extensive simulation experiments using OMNeT++ and compare its performance with a prior motion-driven packet forwarding algorithm. Simulation results show that the SPA can improve the number of delivered packets as well as the average throughput, indicating a viable approach in FANETs.