Flying Ad-Hoc Network Routing Algorithm Based on Knowledge-Graph Embedding

Flying ad-hoc network (FANET) are novel self-organizing networks that have good scalability, robustness, multifunctionality, flexibility, and low cost, making them widely applicable in both military and civilian fields. However, because of the rapid movement of FANET nodes, constant changes in network topology, and limited network resources, developing efficient routing algorithms is challenging. In this paper, we propose a FANET network-topology representation and routing strategy based on knowledge-graph embedding. We model the proposed network to represent nodes, links, and user messages with fixed-dimensional vectors. The model considers node-forwarding delays, link-transmission delays, and network resources to make routing decisions, and uses the routing results of historical messages for further optimization. The simulation results show that the proposed algorithm adapts well to an expanding and dynamic network, reduces the packet-loss rate and end-to-end delay, balances the consumption of network resources, and increases the network lifecycle.