Title: Development of a Data Acquisition System for Unmanned Aerial Vehicle (UAV) System Identification
Speaker: Donald Lear, MS, Aerospace Engineering, & Residential Expert on PX4 UAS Software, ODU
Date: Wednesday, April 25, 2018
Time: 10:00am – 11:00am
Location: NIA, Room 137
Aircraft system identification techniques are developed for fixed wing Unmanned Aerial Vehicles (UAV). The use of a designed flight experiment with measured system inputs/outputs can be used to derive aircraft stability derivatives. This project set out to develop a methodology to support an experiment to model pitch damping in the longitudinal short-period mode of a UAV. A Central Composite Response Surface Design was formed using angle of attack and power levels as factors to test for the pitching moment coefficient response induced by a multistep pitching maneuver.
Selecting a high-quality data acquisition platform was critical to the success of the project. This system was designed to support fixed wing research through the addition of a custom air data vane capable of measuring angle of attack and sideslip, as well as an airspeed sensor. A Pixhawk autopilot system serves as the core and modification of the device firmware allowed for the integration of custom sensors and custom RC channels dedicated to performing system identification maneuvers. Tests were performed on all existing Pixhawk sensors to validate stated uncertainty values. The air data system was calibrated in a low speed wind tunnel and dynamic performance was verified. The assembled system was then installed in a commercially available UAV known as an Air Titan FPV in order to test the Pixhawk’s automated flight maneuvers and determine the final performance of each sensor.
Flight testing showed all the critical sensors produced acceptable data for further research. The Air Titan FPV airframe was found to be very flexible and did not lend itself well to accurate measurement of inertial properties. This realization prohibited the construction of the required math models for longitudinal dynamics. It is recommended that future projects using the developed methods choose an aircraft with a more rigid airframe.
Donald Lear grew up in Cumming, GA, and moved to Asheville, NC to pursue a Bachelor of Science in Physics at the University of North Carolina. In his senior year, he was enrolled in NASA’s Step-Up Internship program, where he developed a MATLAB script to measure spectroscopy data from a pair of orbiting satellites. He then went on to pursue higher education at Old Dominion University as an Aerospace Engineer, where he worked alongside Dr. Drew Landman on various projects in the Unmanned and Autonomous Vehicle Laboratory.
His thesis project focused on creating a data acquisition system for remote controlled (RC) aircraft and using system identification principles to derive said aircraft’s flight characteristics. His work on autopilot systems served to pioneer a new branch of research for the university, and even now he is consulted by other students on similar projects. When he isn’t building Solidworks models and Excel databases, he likes to experiment with new cooking techniques and hone his figure drawing skills