Real-Time Safety-Assured Autonomous Aircraft

ABSTRACT: Aurora Flight Sciences, in collaboration with Worcester Polytechnic Institute, proposes the development of the Real-Time Safety-Assured Autonomous Aircraft (RTS3A) system. The system dynamically performs decision-making based on both sensed and predictive information to carry out adaptive missions and maintenance. This system leverages sensing capabilities distributed throughout subsystems of the aircraft that measure characteristics to allow prediction of the conditional response of the aircraft. The RTS3A system incorporates multi-disciplinary, physics-based models and sensor suites to fully functionalize the flight environment of an aircraft with respect to its structural and propulsion capabilities, which allows for optimization of mission execution as well as condition based maintenance. Distributed sensors and reasoning agents will generate real-time information on subsystems of the vehicle, and each subsystem will communicate with a higher-level system-reasoning agent. A central reasoning agent, informed by the lower level agents as well as real-time updates to the flight environment, will manage mission control systems in order to adapt the maneuver envelop which influences the vehicle"s control authority. The central reasoning agent will ensure minimum margins of flight safety under uncertain and constantly changing conditions. Initial implementation of the self-aware vehicle concept is suggested on unmanned systems, such as Aurora"s Orion Unmanned Aerial System. BENEFIT: This Phase I program will demonstrate the feasibility of a prototype RTS3A system applied to two subsystems, airframe and the propulsion, of a typical aircraft. Prognosis will be based on physics-based models with data to reduce and manage uncertainty. Critical and typical failure modes for the subsystems will be incorporated into the software, allowing aircraft safety to be monitored during uncertain flight conditions. Prognosis of faults and failures in composite airframes allow for the extension of vehicle health management onto the structure, which typically represents the largest portion of the aircraft weight and cost for both maintenance and replacement. Incorporating the propulsion subsystem allows initial realization of the self-aware vehicle concept. Prognostics would allow for condition-based maintenance of airframe and propulsion components, reducing maintenance costs associated with an air vehicle while allowing real-time safety of the vehicle to be monitored.