Autonomous Autorotation Through Expert System Control

The recent rise of autonomous helicopter technology of various scales has led to new demand for automatic control systems that can perform autorotation maneuvers autonomously as part of the flight control system.  Autonomous autorotations have been studied by several investigators, although few have been flight tested and none have been fielded as part of a production avionics system.  Many elements common to previous approaches have limited their utility in practical systems.  Reinforcement learning algorithms require extensive training data that may not be available during control design, and since they are tied to a specific flight platform they are limited in flexibility and scalability to other aircraft.  Iterative optimization algorithms require significant computational capability and, in general, lack convergence guarantees.  New autorotation control laws are needed that provide suitable flexibility and robustness while avoiding iterative optimization, high-fidelity onboard dynamic models, or black-box reinforcement learning approaches.

In this work we have formulated a novel multi-phase expert control system for helicopter autorotation.  The controller is composed of multiple phases which are executed sequentially as the autorotation maneuver progresses.  Phase changes are fuzzily-defined and are initiated through altitude and predicted time-to-impact criteria.  A novel aspect of the control design is its use of time-to-impact predictions during the flare maneuver as well as its use of a desired time-to-impact reference value conditioned on the vehicle’s horizontal velocity, vertical velocity, and rotor kinetic energy.  Simulations results have demonstrated improved performance over that achievable by human pilots, and potential scalability over a broad range of flight platforms.  In several flight experiments in 2013 we flight tested this autorotation controller successfully onboard an Align TREX helicopter.

Plot of autorotation landings Plot of autorotation climb Image of helicopter

References:

  1. Z. Sunberg, J. Rogers, “A Fuzzy Logic-Based Controller for Helicopter Autorotation,” 2013 AIAA Aerospace Sciences Meeting, Grapevine, TX, January 7-10, 2013.
  2. Z. Sunberg, N. Miller, J. Rogers, “A Real Time Expert Control System for Helicopter Autorotation,” American Helicopter Society 70th Annual Forum, Montreal, Canada, May 20-22, 2014.