Because of their small size and agility, quadrotors could revolutionize search-and-rescue or terrain-mapping missions. However, to do so, they have to operate in confined spaces such as rubble corridors or glacial crevasses. Existing models for how quadrotors behave near obstacles are based on helicopter theories, which are inaccurate at the smaller scales of quadrotors. We therefore built a flow-mapping arena to study how micro-scale quadrotors interact with nearby boundaries. We discovered, for example, that dueling vortices appear beneath micro-quadrotors as they land at an angle.
Authors: Darius Carter, Lauren Bouchard, Daniel Quinn
Abstract: The growth of the micro-aerial vehicle (MAV) industry is outpacing our understanding of how MAVs behave in cluttered environments. Search and rescue and product delivery (two key MAV applications) occur in tight, confined spaces filled with complex obstacles. Our current understanding of how micro-quadrotors interact with boundaries is based primarily on helicopter models, which were designed for high-Reynolds-number single-rotor flows. To test how well existing near-boundary models apply to micro-quadrotors, the thrust forces and wakes of a micro-quadrotor near a ground, ceiling, and sidewall were measured. It is found that micro-quadrotors (like their larger counterparts) experience a large boost in lift near the ground/ceiling and a slight drop in lift near the sidewall. Particle image velocimetry is used to quantify the velocity around the rotors and evaluate the assumptions made by existing ground and ceiling models. Complex boundary-layer interactions were observed at low altitudes, especially when the quadrotor was tilted relative to the ground. Reduced-order modeling was also used to explore the safety implications of near ground/ceiling flight. Tradeoffs between safety and efficiency that are sensitive to the ground/ceiling models were discovered, highlighting the need for precise near-boundary models. The results of this study therefore offer guidance for near-boundary model-driven controllers that could improve situational awareness and sensorless landings.