Flying Animal Inspired Behavior-Based Gap-Aiming Autonomous Flight with a Small Unmanned Rotorcraft in a Restricted Maneuverability Environment
MetadataShow full item record
This dissertation research shows a small unmanned rotorcraft system with onboard processing and a vision sensor can produce autonomous, collision-free flight in a restricted maneuverability environment with no a priori knowledge by using a gap-aiming behavior inspired by flying animals. Current approaches to autonomous flight with small unmanned aerial systems (SUAS) concentrate on detecting and explicitly avoiding obstacles. In contrast, biology indicates that birds, bats, and insects do the opposite; they react to open spaces, or gaps in the environment, with a gap_aiming behavior. Using flying animals as inspiration a behavior-based robotics approach is taken to implement and test their observed gap-aiming behavior in three dimensions. Because biological studies were unclear whether the flying animals were reacting to the largest gap perceived, the closest gap perceived, or all of the gaps three approaches for the perceptual schema were explored in simulation: detect_closest_gap, detect_largest_gap, and detect_all_gaps. The result of these simulations was used in a proof-of-concept implementation on a 3DRobotics Solo quadrotor platform in an environment designed to represent the navigational diffi- culties found inside a restricted maneuverability environment. The motor schema is implemented with an artificial potential field to produce the action of aiming to the center of the gap. Through two sets of field trials totaling fifteen flights conducted with a small unmanned quadrotor, the gap-aiming behavior observed in flying animals is shown to produce repeatable autonomous, collision-free flight in a restricted maneuverability environment. Additionally, using the distance from the starting location to perceived gaps, the horizontal and vertical distance traveled, and the distance from the center of the gap during traversal the implementation of the gap selection approach performs as intended, the three-dimensional movement produced by the motor schema and the accuracy of the motor schema are shown, respectively. This gap-aiming behavior provides the robotics community with the first known implementation of autonomous, collision-free flight on a small unmanned quadrotor without explicit obstacle detection and avoidance as seen with current implementations. Additionally, the testing environment described by quantitative metrics provides a benchmark for autonomous SUAS flight testing in confined environments. Finally, the success of the autonomous collision-free flight implementation on a small unmanned rotorcraft and field tested in a restricted maneuverability environment could have important societal impact in both the public and private sectors.
Sarmiento, Traci Ann (2017). Flying Animal Inspired Behavior-Based Gap-Aiming Autonomous Flight with a Small Unmanned Rotorcraft in a Restricted Maneuverability Environment. Doctoral dissertation, Texas A & M University. Available electronically from