Development and Flight Testing of a Model Based Autopilot for a Low Cost Unmanned Aerial System

The availability and proliferation of small unmanned aircraft systems has provided engineers with the opportunity to gain hands on experience with aircraft in a manner that was previously limited to institutions and companies with large budgets. Even though the scope of these development programs are at their minimum, it is still challenging to simultaneously learn the skills required to develop an autonomous air vehicle while also applying the lessons to a physical system. In addition, current systems typically require a multidisciplinary team of engineering types to be successful in a short academic term. In this paper, a method for streamlining the software and hardware development of an unmanned aerial vehicle is developed and allows engineers with no specific background in electrical or computer engineering to leverage the power of Matlab®/Simulink® to develop, simulate, and ultimately flight test a system within an academic term using a very low cost autopilot system. Specific focus has been spent during the development of this tool on obfuscating the computer and electrical sciences from the problem so that engineers may concentrate on the more abstract concepts relevant to their field of study. In addition to a detailed description regarding the development of these tools, this paper presents simulation and flight testing results of the integrated system. An LQR based control law was successfully implemented and demonstrated in flight ultimately allowing the aircraft to navigate a course of waypoints autonomously. As a result of the successful development, the tool has been made available in the public domain for widespread distribution and hopeful collaboration with other aspiring developers.