Yazar "Hasseni, Seif-El-Islam" seçeneğine göre listele

Listeleniyor 1 - 2 / 2
  • Küçük Resim Yok
    Yayın
    Genetic algorithm optimization of quadcopter PID control
    (Maltepe Üniversitesi, 2018) Hasseni, Seif-El-Islam; Başoğlu, İsmail; Bayülgen, Batu; Öztopcu, Aslı; Salman, Ayşe
    The Quadcopter control makes a challenge for researchers, because it is a nonlinear system, which is complex and strongly coupled. lt is at 6 degrees of freedom; 3 for rotation and 3 for translation, but only 4 inputs, what makes it an under-actuated system. Many non-linear control approaches were applied to this kind of systems; among them, we can find sliding mode, back-stepping and other techniques with different complexity design.
  • Küçük Resim Yok
    Yayın
    Genetic algorithm optimization of Quadcopter Pid Control
    (Maltepe Üniversitesi, 2018) Hasseni, Seif-El-Islam
    The Quadcopter control makes a challenge for researchers, because it is a nonlinear system, which is complex and strongly coupled. It is at 6 degrees of freedom; 3 for rotation and 3 for translation, but only 4 inputs, what makes it an under-actuated system. Many non-linear control approaches were applied to this kind of systems; among them, we can find sliding mode, backstepping and other techniques with different complexity design. Linear approaches; were also applied for non-linear systems, but the linearization is required which leads to the fact that the application will be achieved just for rotational subsystem with a small range of angles. In this work, the PID controller is proposed for the full system and so, to keep the nonlinearity, we have used a meta-heuristic optimization technique to tune the parameters, based on a Genetic Algorithm.Our strategy is as follows: a. Designing the controllers of the actuated subsystems (altitude and attitude) by a decentralized way. Here the interactions between the altitude and the angles are decoupled by adding an adaptive-term. b. Decoupling the interactions between the translation and the angles by applying a sequential design; The controllers are designed for the actuated system and then the position controllers by a cascade design. The position coordinates have virtual inputs, which are the accelerations. The use of GA allows us to design a simple linear control low for a complex nonlinear system. The proposed technique is not limited to the only case of linear range and is extended to a large range of states; this is not possible if we use classical approaches. There are output states (translation position) with under-actuated degrees of freedom, and which have a virtual inputs; our method, contrary to the classical ones, has the ability to design the controllers by using a Genetic Algorithm which allow giving the optimal parameters. The simulations are implemented in MATLAB/ Simulink tool to evaluate the control technique in terms of dynamic performance and stability. In spite of the simplicity of the PID controllers’ design, the effect of the proposed technique is shown in terms of tracking errors and stability, even for big angles, subsequently, high velocity response, high dynamic performances, and practically acceptable rotors speed.