Implementation of Biomimetic Precision Flight in Autonomous Air Vehicles

Physical Sciences Inc. (PSI) and the University of Arizona (UA) propose to investigate, evaluate, and apply the sensory schemes, neural processing, and behavioral strategies of flying insects to the problem of maneuvering flight of an Unmanned Air Vehicle(UAV). We will develop computational models of vision (and other sensors) and insect flight control processes incorporating critical sensor-motor neural pathways. A comprehensive insect vision model, based on fundamental structural biology data andevolutionary biology, will be developed and calibrated against cell response measurements. The structural and electrophysiology work underpinning this model will be performed at UA. PSI will model pitch/roll, yaw, forward acceleration and attitude controlprocesses, relating sensory inputs to motor neuron outputs (i.e., actuator inputs). These models, collectively called the flight dynamcis model, will be correlated to available behavioral data on insect flight. The dynamics model will then be used togenerate guidance inputs (course correction commands) for a UAV flight simulation model. This coupled simulation will assess the performance of a UAV with insect-like vision (and possibly other) sensors in performing maneuvers such as obstacle avoidanceand target tracking. A breadboard model of a key component/subsystem of insect vision or control processors will also be sonstructed by PSI.