Conference article

The two degrees of freedom (2-DOF) helicopter is an openloop unstable multi-variable process. Various control strategies can be applied to stabilize the system for tracking and regulation problems but not all control methods show equal capabilities for stabilizing the system. This paper compares the implementation of a classical PID controller, a linear quadratic regulator with integral action (LQR+I) and a model predictive controller (MPC) for stabilizing the system. It has been hypothesized that for such an unstable MIMO (multi input multi output) process showing cross coupling behavior, the model based controllers produces smoother control inputs than the classical controller. The paper also discusses the necessity of including the derivative part of the PID controller for stabilization and its influence to the measurement noises. A Kalman filter used for estimating unmeasured states may produce bias due to model mismatch. The implementation and comparison is based on a 2-DOF experimental helicopter prototype.

2-DOF helicopter, MPC, LQR, PID, Kalman filter, qpOASES