Route Selection in 5G-based Flying Ad-hoc Networks using Reinforcement Learning
Muhammad Fahad Khan, Kok‐Lim Alvin Yau
Abstract
Flying ad-hoc network (FANET) is one of the applications of next-generation wireless networks, including fifth generation (5G) networks. Due to the availability of high data rate and low latency, 5G supports applications with high resource requirement including FANET. In FANETs, unmanned aerial vehicles (UAVs) communicate with each other to form a network, however due to their high mobility and dynamicity, network topology varies frequently. UAVs have limited energy resources, and their energy is consumed for communication, such as data transmission and reception. In this paper, the optimal route is identified by considering UAVs with higher residual energy and stability in a 5G network in order to increase network lifetime, as well as reduce energy consumption and the number of broken links. The proposed algorithm applies an artificial intelligence approach called reinforcement learning, and the effects of different learning rates are investigated.