This paper investigates traffic flow modeling issue in multi-services oriented unmanned aerial vehicle (UAV)-enabled wireless networks, which is critical for supporting future various applications of such networks. We propose a general traffic flow model for multi-services oriented UAV-enable wireless networks. Under this model, we first classify the network services into three subsets: telemetry, Internet of Things (IoT), and streaming data. Based on the Pareto distribution, we then partition all UAVs into three subgroups with different network usage. We further determine the number of packets for different network services and total data size according to the packet arrival rate for the nine segments, each of which represents one map relationship between a subset of services and a subgroup of UAVs. Simulation results are provided to illustrate that the number of packets and the data size predicted by our traffic model can well match with these under a real scenario.

Unmanned Aerial Vehicles (UAVs) are opening up new opportunities for extensive applications. The traffic flow model is critical to evaluate the traffic needs of various applications in designing and deploying UAV system. However, the traffic flow model has not been explored in multi-services oriented UAV system. To this end, this paper proposes a general traffic flow model for multi-services orientated UAV system. Under such a model, the network services are first categorized into three subsets, each corresponding to one of telemetry, Internet of Things (IoT), and streaming data. According to the Pareto distribution, all UAVs are further partitioned into three subgroups relying on their network usages. We can measure the packet arrival rate for the nine segments, each of which represents one map relationship between a services subset and a UAV subgroup. Therefore, we can also obtain the number of packets for each network service and total data size. Simulation results are presented to illustrate that the number of packets and the data size predicted by our traffic model can well match with these in real scenarios with different network services.