Abstract:
Technological advancements in the field of computers have urged digital implementation of
control techniques. Sampled data control of quadrotor unmanned aerial vehicle (UAV) is
presented in this document. Two control techniques, sliding mode control (SMC) and
backstepping (BS) have been applied in this regard to stabilize the quadrotor UAV. And for each
technique four discrete time (DT) controllers are extracted, two from the discretization of the
continuous time (CT) control based on continuous time system model (CSTM) and two
controllers are based on approximate discrete time equivalent system model (ADTESM). Full
state stabilization of quadrotor UAV has been achieved in a semiglobal practical sense.
ADTESM of quadrotor has been obtained via Euler forward difference (EFD) method. In order
to establish the robustness of the controllers Monte Carlo simulations have been performed. And
these simulations have revealed that DT controllers based on ADTESM are far more robust than
the DT controllers based on CTSM. Simulations for different initial conditions have revealed
that the controllers designed based on ADTESM have larger regions of attraction as compared to
the controllers obtained by discretizing the CT controls based on CTSM. Simulations in the
presence of the wind gust have also been carried out and a comparative analysis has been
presented as well.
The proposed Discrete time sliding mode control (DTSMC) with and without integral action not
only addresses the issue of chattering that is inherited in SMC but also achieves full state
stabilization of quadrotor UAV in a semiglobal practical sense. Additionally, proposed discrete
time backstepping control (DTBSC) with and without integral action based on ADTESM
robustly stabilizes the UAV with parameter perturbations of ± 30 percent randomly generated
tolerance. While the DTBSC based on CTSM destabilizes even with the parameter perturbations
of ± 3 percent tolerance and discrete time integral backstepping control (DTIBSC) based on
CTSM can withstand up to ± 5 percent parameter perturbation in random
