Model Predictive Control on a Fully-actuated Octocopter for Wind Disturbance Rejection

Caleb Probine, David Yang, Karl A. Stol, Nicholas Kay

Abstract:

Physical interaction with the outdoors is a key
application of UAVs. However, in these situations aircraft must compensate for external disturbances such as wind. Fully-actuated UAVs are
well suited to this, but consideration is needed
in control design. This work focuses on how
Model Predictive Control (MPC) can be applied
here. MPC has the advantage of being able to
take system constraints into account, and its utility is investigated in wind rejection tasks. This
work has focused on the formulation of appropriate constraints on craft-orientation and motor
efforts. Also incorporated are frequency dependent actuator weightings, and basic disturbance
estimation. Overall, in simulation, modelling of
motor constraints was not found to be effective
when compared to prior controllers. However,
the use of MPC with angle constraints was found
to improve position regulation. The controller
was then verified with wind tunnel experiments.