Note:

Abstract: We consider the problem of re-initialization or attitude stabilization of a Quad-rotor under large impulse disturbances or arbitrary initial conditions and address it using the framework of quasi-Linear Parameter Varying (qLPV) systems applied to the attitude dynamics represented in quaternion coordinates.Due to the absence of any approximations or linearization errors, we show in simulation that large disturbances can be well rejected and the quad-rotor can be stabilized from arbitrary initial conditions which is not possible by using a linearized model.Furthermore, we show that the qLPV representation obtained in the quaternion coordinates is affine and that a gridding based controller tuned for a sufficiently coarse grid outperforms the LTI controller.

Note:

Abstract: We consider the problem of re-initialization or attitude stabilization of a Quad-rotor under large impulse disturbances or arbitrary initial conditions and address it using the framework of quasi-Linear Parameter Varying (qLPV) systems applied to the attitude dynamics represented in quaternion coordinates.Due to the absence of any approximations or linearization errors, we show in simulation that large disturbances can be well rejected and the quad-rotor can be stabilized from arbitrary initial conditions which is not possible by using a linearized model.Furthermore, we show that the qLPV representation obtained in the quaternion coordinates is affine and that a gridding based controller tuned for a sufficiently coarse grid outperforms the LTI controller.