Autonomous Autorotation of an Unmanned Helicopter Using a Reinforcement Learning Algorithm

Abstract

Single-engine manned helicopters can face unexpected engine failure during their mission operations. In this situation, pilots typically perform an autorotation maneuver to a safe landing, which demands great skill on the part of the pilot. Similarly, unmanned helicopters can encounter engine failure during autonomous operations. Typically, an autorotation maneuver of a helicopter under engine failure can be divided into three phases. First, the entry phase consists of controlling the angular motion of the helicopter so that the main-rotor rotation decays. Second, during the steady-state descent phase, air flows upward through the rotor disk. Potential energy of the helicopter is converted to kinetic energy to attain desired steady-state descent airspeed below the maximum sink rate. In the final phase, the airspeed and the sink rate are reduced just before touchdown.