Realistic models for characterizing the performance of unmanned aerial vehicles

Unmanned Aerial Vehicles (UAVs) are increasingly being adopted for military and civilian applications. UAVs available on the market are well known to be resource constrained, especially in terms of available energy. As a result, it is very challenging to predict the critical performance characteristics of a UAV, such as flight time or the ability of a UAV to complete a mission, given the system parameters. Nevertheless, such predictions would have several benefits, such as improving the effectiveness of mission planners and optimization algorithms in general, as well as enabling researchers to perform more realistic simulations. The goal of this paper is to gain understanding in how physical, mechanical, or electrical hardware aspects of a UAV affect the UAV performance and ultimately its capability to accomplish a mission. We propose two models for UAV performance. The first model considers basic UAV operations, while the second model considers the UAV physical characteristics as well as the mission specifications to predict the UAV flight time and the number of waypoints it can safely traverse. We validate our models thorough experiments using a real test-bed based on 3DR Solo UAVs. The results show that our approach is able to reliably predict performance to within less than 5% margin of error.