Implementation of the network-based moving sliding mode control algorithm to the rotary inverted pendulum system

Abstract

In this study, the most preferred in control applications rotary inverted pendulum system (rip) is dealt. The coordinates of the center of gravity of the rip elements were found and the total kinetic and potential energies of the system were obtained. Lagrange function has been formed by using kinetic and potential energy expressions. Expressions giving the equations of motion of the system have been found by taking into consideration the Lagrange method. Using the state variables, the pendulum angle of the system has been controlled by the moving sliding mode control method via the program written in Matlab. The slope of the sliding surface is calculated by artificial neural networks. Optimum values of weight and bias coefficients of artificial neural networks are found by using the genetic algorithm. From the results, it has been seen that the pendulum angle reaches to about 25 Nm motor torque and the reference value reaches about 3 seconds and the error is about zero.