A Dynamic and Local Perspective: A Reliable Q-Routing Protocol based on Fuzzy Logic for FANETs.
Penghui Hu,Jian Wang,Yirui Huang,Yan Ma,Cunzhuo Zhao,Chaofan He
DOI: https://doi.org/10.1109/ICCC57788.2023.10233540
2023-01-01
Abstract:Flying Ad-hoc Networks (FANETs) have found extensive applications in various fields, such as agriculture, security, rescue operations, and military or civilian operations. However, ensuring reliable routing and energy efficiency remains a critical challenge in FANETs due to the high mobility of drones and limited wireless resources. This paper investigates a general routing problem, specifically how battery-limited UAVs can route data reliably over multiple hops using only local information. We propose a multi-hop exploration model based on the local time-varying graph. We highlight the diminishing information freshness during hop-by-hop exploration, which limits the perceived validity of the network topology. This limitation often leads to unstable transmission when making routing decisions. To maximize routing reliability in long-term exploration, we formulate an appropriate optimization problem. Additionally, we introduce reinforcement learning and fuzzy logic into the routing protocol design, named FQR. To evaluate the quality of wireless links using a lightweight solution, we employ a fuzzy logic approach that considers multiple metrics, including relative mobility, resident energy, and link quality, as reliability features. We enhance the exploration process by incorporating information from two-hop neighbors in the reward definition. Furthermore, we design a variable discount factor to differentiate the validity of different local links. The proposed method is evaluated through simulations in NS3, utilizing an actual UAV trajectory model. The results demonstrate that our approach outperforms traditional strategies regarding packet delivery ratio, end-to-end delay, and network lifetime.