Mohammad Farahmand, Reza Ghasemi, Mohammad Salari,
Volume 18, Issue 1 (3-2018)
Abstract
Controller design for non-linear multi-input, multi-output systems, such as unmanned quadrotor vehicles, has always been a challenging issue due to the strong interconnection between state variables and highly nonlinear dynamic equations. In addition, quadrotor is an under-actuated non-linear dynamic device. Due to being under-actuated for moving in the horizontal direction, the combination of changes in the speed of the existing quadruple operators should be used. So that, by creating the angle between the quadrotor hypothetical plane and the horizon surface, the device can be forced to move in the longitudinal or transverse direction. Therefore, in the quadrotor control system, two nested control loops are required. An outer loop to determine the appropriate angle of the device relative to the horizon for horizontal movements and an inner loop that is required to angle of the device panel is equal to this angle. In this paper, a fuzzy hybrid super-twisting sliding mode non-linear controller for controlling a sample quadrotor is designed. For this purpose, a fuzzy controller in the outer loop and a super twisting sliding mode controller in inner loop are used. An important advantage of this strategy is that it optimizes the horizontal speed of the device. If the distance from the target is too high, the angle of the device panel also increases, and if the distance is reduced, the angle also decreases. As a result, the device reaches the target with the desired speed. The performed simulation results confirmed this fact.
M. Aalipour, A. Mokhtarian, H. Karimpour,
Volume 20, Issue 3 (2-2020)
Abstract
Spherical robots are the mobile robots with spherical shapes equipped to an internal drive mechanism that moves on the ground due to their external shell rolling. In this research, first, a pendulum spherical robot is modeled, then using the Lagrange method, dynamic equations of plane motion of robot on the non-flat surface are derived. Considering the scarcity of the number of operators relative to the number of degrees of freedom of the spherical robot, designing of a non-linear controller is performed based on feedback linearization techniques. Therefore, regarding non-confirm initial conditions on the trajectory, parametric uncertainty and disturbance torque on the robot, the performance of the system has been investigated. By selecting the appropriate rotation trajectory, the robot motion is simulated in MATLAB software and in following the pendulum rotation angle and actuating torque are obtained. The results indicate that the designed controller has proper and resistant performance in tracking selected trajectory for sphere shell rotation during moving on a non-flat surface.
