Switching controller design to swing-up a two-link underactuated robot

Abstract

In the control engineering field it has shown that the dynamical behaviors are mostly non-linear. Meanwhile, the solution for facing these challenges seem to be not feasible or despite simulations, they just don't come to the good results. Therefore, in this research a practical approach is used which can lead a system to any operating point without considering its region of attraction. In other words, this solution presents a way for finding a path which brings a two-link robot system from one equilibrium point to another. Two-link pendulums are widely used as an example of highly nonlinear underactuated mechanical systems. The limited working space of an inverted two-link pendulum despite other mechanical systems such as cart and pendulum system requires some consideration for the controller design. Actually, in a cart and pendulum system one of its degree of freedom (the linear position of the cart) may vary from minus infinity to plus infinity which is a great advantage for controller design but such a thing does not exist in the case of two-link inverted pendulum. By using the proposed method, a swing-up controller has been successfully employed. In order to control the two-link robot, feedforward control and feedback control have been used. The feedforward control is achieved by Discrete Mechanics and Optimal Control (DMOC) method. In order to optimize the path of objective, Hamiltonian method has been employed. As a verification of this method, experimental and simulation results have been compared. Regarding the implementation of the two-link robot, due to the lack of accuracy in dynamic parameters, the feedback control has been used.