2DOF Helicopter Models: a simulation evaluation for MPC applications

Model Predictive Control (MPC) is a well-known technique, able to guarantee acceptable performance by anticipating future output trends and taking control actions accordingly. In the most common case, it is used to control slow (chemical) processes. Therefore, a challenging case of application is nonlinear Unmanned Aerial Vehicles (UAVs), requiring sampling times under 20ms. In this work, different two-degree-of freedom (20F) helicopter models are evaluated within the context of MPC applications. The proposed models are compared in terms of run time, prediction error and closed-loop control performance. It is concluded that in this case, it is sufficient to use a single Linear Time Invariant (LTI) model for prediction since the closed-loop control performance does not increase considerably using more complex models.