Yazar "Ozsoy, Can" seçeneğine göre listele

Listeleniyor 1 - 3 / 3
  • Küçük Resim Yok
    Öğe
    Design of MPCs for a fixed wing UAV
    (Emerald Group Publishing Ltd, 2017) Ulker, Hakan; Baykara, Cemal; Ozsoy, Can
    Purpose - The purpose of the paper is to examine a fixed wing unmanned aerial vehicle (UAV) as it undergoes five flight scenarios under straight and level, level climb, level turn, climbing turn and level steady heading sideslip conditions in a desired and controlled manner using constrained multi input multi output (MIMO) model predictive controllers (MPCs). Design/methodology/approach - An MPC strategy based on the lateral and longitudinal linear models is proposed for the flight control design. Simulations are carried out for the nonlinear closed-loop aircraft Simulink model available from the University of Minnesota UAV research group with the implemented MIMO MPCs designed in this paper. Findings - The results of the simulations show that the MIMO MPCs can achieve satisfactory performance and flying qualities under three different test conditions in terms of existing unmeasured outputs and unmeasured output disturbances. Practical Implications - The MPCs designed in this paper can be implemented to UAVs. Therefore, the implementation is considered as an advanced research. Originality/value - The proposed MPC design in this paper provides more flexibility in terms of tracking complex trajectories comparing with the classical controllers in the literature. Besides they provide to change more than one reference of the states at any time.
  • Küçük Resim Yok
    Öğe
    Hovering stability of a model helicopter
    (Emerald Group Publishing Ltd, 2016) Koc, Ilker Murat; Franko, Semuel; Ozsoy, Can
    Purpose - The purpose of this paper is to investigate the stability of a small scale six-degree-of-freedom nonlinear helicopter model at translator velocities and angular displacements while it is transiting to hover with different initial conditions. Design/methodology/approach - In this study, model predictive controller and linear quadratic regulator are designed and compared within each other for the stabilization of the open loop unstable nonlinear helicopter model. Findings - This study shows that the helicopter is able to reach to the desired target with good robustness, low control effort and small steady-state error under disturbances such as parameter uncertainties, mistuned controller. Originality/value - The purpose of using model predictive control for three axes of the autopilot is to decrease the control effort and to make the close-loop system insensitive against modeling uncertainties.
  • Küçük Resim Yok
    Öğe
    PIL simulations of an FWUAV under windy conditions
    (Emerald Group Publishing Ltd, 2018) Ulker, Hakan; Baykara, Cemal; Ozsoy, Can
    Purpose A fixed wing unmanned aerial vehicle (FWUAV) is targeted to perform processor in the loop (PIL) simulations for the flight scenarios such as straight and level, level climb, level turn, climbing turn and level steady heading sideslip under windy conditions such as steady wind (SW) and wind gust (WG) in a desired and controlled manner. Design/methodology/approach The constrained multi-input-multi-output (MIMO) lateral and longitudinal linear models-based model predictive controllers (MPCs) which are proposed in a previous study (Ulker et al., 2017) are tested in the PIL simulations under specified windy conditions. BeagleBone Black Rev C is used as a target hardware or processor in the PIL simulations. Findings The results of the PIL simulations show that the MPCs proposed in the previous study can achieve satisfactory performance and flying qualities for the all flight scenarios handled in this paper under windy conditions. Practical implications The MPCs proposed in the previous study can be easily implemented in the real world to a low-cost and small-sized board like BeagleBone Black Rev C which is used in this paper. Originality/value The proposed MPCs in the previous study which are capable of providing more flexibility in terms of tracking complex trajectories are showed to be able to be implemented to real system by means of PIL simulations under the changeable windy conditions which are difficult for performance tests.