Drag Characterization of a Fixed-Wing Unmanned Aerial Vehicle (UAV) with COTS Avionics Through Sink Tests

An automated flight testing system was developed to perform steady state and dynamic measurements of key parameters to characterize the aerodynamic performance of small unmanned aerial vehicles (UAVs) using commercial off-the-shelf (COTS) components. These include the parasitic drag coefficient, Oswald efficiency, and various stability parameters necessary to characterize the dynamic modes of motion of an airplane. We demonstrate the modularity of the system by performing flight tests with two different aircraft. Aerodynamic testing of a novel UAV requires flying repetitive flight profiles at specific flight conditions, possibly for an extended period of time to allow for sound measurements. Maintaining a sufficiently steady condition necessary for these tests would be impractically difficult to conduct with line-of-sight, remote-controlled piloting. To that end, we develop a COTS avionics platform to autonomously fly such profiles using an Ardupilot-based controller with pre-programmed missions designed to sample specific coefficients. To determine static coefficients, namely parasitic drag coefficient and Oswald efficiency, we perform a constant-airspeed sink test. The controller trims the aircraft at a constant airspeed at idle power, and measures the sink rate. The test is repeated at various airspeeds to generate a sink polar. The static coefficients can then be calculated as functions of the coefficients of a quadratic-fit over the polar data. To determine the dynamic longitudinal response of the aircraft, we perform doublet tests. The controller trims the aircraft at a constant airspeed and throttle. The elevator is given a step input in both directions, exciting a longitudinal mode of motion. The controller gives up control while measuring a time-domain history of Euler angles and rates, body angles, and accelerations using the onboard IMUs and accelerometers. The data can be used to calculate the natural frequency, damped frequency, and damping ratio of the response with classic System Identification techniques, from which second-order stability parameters can be determined. The system is designed to be modular such that any UAV can be characterized with the profiles flown by the controller. To demonstrate such modularity, we perform autonomous flight test routines with two different aircraft: a custom-designed-and-built 10ft wingspan, high aspect ratio conventional configuration foam aircraft, as well as a commercially-sold 3ft wingspan small UAV flying wing. Though designed for modularity and to be applicable to any UAV, the system was developed as part of the Human-Powered Flight Research Team at USC, a newly founded group with the purpose of beating the endurance world-record of a human-powered aircraft, set by MIT in 1988. To that end, HPFRT is developing several platforms to serve as technology demonstrators for the eventual construction and test flights of the full-scale, 100+ ft wingspan aircraft. One of these platforms is a quarter-scale aircraft. Its main purpose is to validate the sizing codes, aerodynamic models, and construction techniques for the full-scale aircraft. The quarter-scale demonstrator is expected to take flight around May of 2025. A crucial part of the quarter-scale program is the collection of high-quality data for full characterization of the aircraft’s static coefficients and dynamic modes. This is unprecedented at the student design team level. To that end, we develop the Human-Powered Aircraft Avionics Testbed (HATB) UAV. This is a high aspect-ratio aircraft - 10 ft wingspan, 1 ft constant chord - designed to cruise at a similar lift coefficient as a human-powered aircraft with the DAE-31 airfoil used for the wing. It is designed to fly at 11 m/s, and weighs close to 10lb. The avionics package was tested on this aircraft, as well as the small, commercially available flying wing.