Development of a software in the loop simulation approach for risk mitigation in unmanned aerial system development

Abstract

A common method to reduce risk during the development of new designs is simulation and estimation. The extent to which these simulation and estimation techniques can be relied upon for small Unmanned Aerial Systems (sUAS) is unknown. Combining the autopilot together with a simulator for Software in the Loop (SITL) allows designers to tune and observe autopilot behavior before the design is finished. This thesis extends the tools provided through SITL to present methodology that can both provide early insight to the handling qualities of the aircraft and validation of the simulator.

ArduPilot and X-Plane 11 are used as the autopilot and simulator. New features were developed to extend the functionality of ArduPilot. Additional software was developed to assist in both identification of aircraft modes and validation of simulation. With mixed results from validation of X-Plane, the need to perform flight testing of real aircraft is still more desirable for precision tuning of sUAS for more desirable handling qualities. What can be gained from SITL is risk mitigation from unconventional additions to sUAS and detailed analysis of failsafe behavior of the autopilot.