Micro Aerial Vehicles (MAVs) have the potential to revolutionize search-and-rescue, product delivery, aerial mapping, and weather sensing. Current models for how MAVs interact with solid boundaries are based on helicopters, which operate at different length and speed scales. Here we compare existing and updated models with experimental thrust and flow field measurements, and we demonstrate how better near-ceiling and near-ground models could improve MAV efficiency and situational awareness. (This work was done in collaboration with the Bezzo Lab at the University of Virginia.)
Authors: Shijie Gao, Carmelo Di Franco, Darius Carter, Daniel Quinn, Nicola Bezzo
Abstract: Micro Aerial Vehicles (MAVs) and in particular quadrotors have gained a lot of attention because of their small size, stable, robust, and diverse sensing capabilities that make them perfect test beds in several safety critical operations. Shrinking these vehicles is desirable since agility increases. However, it entails smaller power sources and hence less flight time. Adding sensors on these systems also implies more energy consumption due to both the added weight and the supplied energy to the sensors. In this work, we build a framework to leverage the flow dynamic effects near surfaces to recognize grounds and ceilings during operations and to plan a trajectory while minimizing energy consumption. Our proposed framework leverages data from real experiments to model the behavior of the system near surfaces and graph theoretical approaches for energy efficient motion planning. As a result, this study indicates that i) we can detect surfaces during operations without the need of extra onboard sensors and ii) we can minimize energy consumption up to 15% when the system can fly near ground or ceiling surfaces. The proposed framework is validated with experimental results on a quadrotor UAV.