Network Coded Rate Decay Flooding: Position Aware Network Coding in Large-Scale Urban UAV Networks

In urban areas, the extensive use of Unmanned Aerial Vehicles (UAVs) promises to revolutionize many industries but will increase the number of UAVs inhabiting a given airspace drastically. A necessity for this is the establishment of an Unmanned Aerial Traffic Management (UTM) system that enables the safe and reliable operation of UAVs. One way to establish this is to form an ad-hoc network between UAVs, operators and ground stations and distribute movements and commands throughout the entire network. Recently, Rate Decay Flooding (RDF) was proposed to realize such a network by gradually delaying packets the further they travel and thereby reducing the traffic density. Furthermore, Network Coding (NC) is a novel approach to reduce the traffic as different pieces of information can be encoded into a single transmission. In this work, we present a combination of both approaches. For this we propose a coding policy that determines when to use NC based on the contents of packets waiting to be forwarded. Further we implement a Bayesian online estimator for the current packet loss rate, which is a crucial component of the coding policy and thereby allowing the protocol to adapt to changing network sizes. The protocol is evaluated in an open-source simulator. Here we show that the coding policy successfully determines high-gain coding actions and reacts to larger networks with higher packet loss, by reducing the share of coded transmissions. Using our approach networks 95 % larger than previously possible with RDF are supported by reaching 21.2 % more receivers per transmission.