A Fountain Code Enhanced Geographical Multipath Routing Strategy for UAV Networks

The burgeoning deployment of unmanned aerial vehicles (UAVs) in dynamic environments to perform complicated tasks has engendered significant challenges in UAV networks, predominantly due to unstable communication links and changing network topology. This paper proposes a distributed geographic multipath routing strategy based on fountain code, achieving low-latency and reliable routing under dynamic and complex UAV network topologies. We conduct the theoretical analysis of end-to-end latency based on the relationship of distinct encoded packets among the UAVs, and propose an joint optimization model to enhance routing efficiency. Since that the centralized routing with overall joint optimization is not adapt the dynamic topology, we propose a distributed approach based on geographical multipath routing. This approach divides the original problem into multiple sub-problems, which are executed separately at the source UAV and intermediate UAVs, thus achieving distributed routing decisions. Unlike existing work that ignore the differences of encoded packets, we further introduced a cross-entropy-based transmission control model that enables the forwarding of distinct packet groups. Simulation experiments indicate that this solution can effectively reduce overall communication latency and improve route reachability.