Studying the interaction of UAS and human pilots using the X-Plane flight simulator

Abstract

A system of networked computer components is described allowing realistic simulation and performance assessment of see-and-avoid algorithms for unmanned aircraft systems (UAS). Due to its realistic physics engine and relatively open architecture, the X-Plane flight simulator is adopted to simulate computer-generated traffic or an aircraft controlled by a human pilot. Using the UDP interface in X-Plane, real-time collision avoidance algorithms are then implemented, where the simulated unmanned aerial vehicle (UAV) controller reads aircraft positions from X-Plane and updates control inputs to the UAV using sequential quadratic programming (SQP) to avoid collisions in real time. The basic concept is illustrated through an example flight involving a single aircraft in an airport pattern and a UAV flying through the same airspace. A second experiment is performed where a human pilot intentionally flies across the path of the UAV.