Application of Software Enabled Control Technologies to a Full-Scale Unmanned Helicopter

This paper presents a control architecture designed to accommodate a selection of modern control algorithms on a full-scale rotary-wing, unmanned aerial vehicle. The architecture integrates a visual landing system, two path planners, a flight envelope protection algorithm, and two low-level flight controllers that were developed independently by six agencies in academia and industry. A newly developed vehicle model and an exportable simulation environment were assembled in an open control infrastructure to expedite the algorithm development. The collaboration resulted in successful flight testing of the architecture and multiple control algorithms on Boeing’s Renegade Unmanned Aerial Vehicle, a derivative of the Robinson R22. The aircraft successfully switched from a conventional flight controller to an adaptive neural network flight controller on four occasions making it the largest helicopter to operate under adaptive neural network flight control.