Multi-Source Markovian Model for Video Streaming in Flying Ad hoc Networks

Abstract

In multi-drone networks, video streaming helps in performing many tasks such as surveillance, inspection, and map generation. In such networks, packet loss frequently occurs due to the high mobility of the nodes and to congestion. Therefore, a model is required in order to investigate and analyze the trade-off between the end-to-end delay, packet dropping probability, and the required number of paths. In this paper, multi-source live-video streaming in Flying Ad Hoc Networks FANETs is analyzed, and a network queue model is suggested. Then, the quality of the paths is estimated, experiments are conducted, and conclusions based on the results are obtained. The network is modelled with an M/M/1 Markovian model to quantify the system’s performance. Transition sequence between good and bad states is taken into account by the proposed model as a packet loss process. The good and bad stationary probability is used, after measuring the Packet Reception Ratio PRR, in order to discriminate the good paths from the weak ones. This multi-source approach increases the network goodput and allows the system to be unlimited by the weakest source node quality.