Autonomous flight with hierarchical planning
摘要
A hierarchical planning method with three levels was demonstrated in simulation and in autonomous flight on a Black Hawk helicopter. The reactive terrain avoidance software has been previously flown and has demonstrated the ability to sense and avoid previously unknown terrain while autonomously guiding the aircraft to a destination. The new mission level software adds the ability to make mission-length plans based on a stored terrain map while minimizing mission risk. Pilot or operator input can be as minimal as simply entering a destination and waiting, but the pilot or operator also has the capability to re-plan while en route, change speed, and add direction bias to influence the flight path. The system was flight-tested at four different sites with different types of terrain. The addition of this mission level improves the ability of the system to keep to a desired height above ground level, while allowing the pilot or operator to better predict the route of the aircraft and time to destination.