SQBRP-SDFANET: A scalable Q-learning-based routing protocol for SD-FANETs

Abstract

In today’s fast-evolving world of wireless networking, creating efficient and reliable communication systems is essential. In this context, we introduce a new Q-learning-based routing protocol designed specifically for flying ad hoc networks (FANETs) that enhances path selection and network performance. Our protocol tackles the unique challenges of FANETs, such as dealing with rapidly changing topologies and managing UAV resources. We achieve this objective through various techniques, including partitioning the area of interest into hexagonal cells, reducing the exploration space to a specific angle, and enabling recovery in case of failure. The protocol scales effectively through geographic partitioning, as the size of the Q-learning table is determined by the number of hexagonal cells rather than by the number of UAVs. Path calculation is performed in two stages: initially, a path composed of cells, and then a mapping that represents each cell as a node with the best relay characteristics. Extensive simulation for dense network has been driven under different conditions, including varying cell sizes, UAV densities and speeds, network load, and learning period to demonstrate the robustness of our protocol in terms of packet delivery ratio and transmission delays.