Regresar

Identification of a Ball-Plate System Using Ant Colony Algorithm

Abstract:

The objective of the ball-plate system is to control the position of a spherical body on a plane surface by means of its angular movement, that is to say, from the plane's rotations, the movements of the sphere are controlled, involving several magnitudes such as mass, position, velocity, inertia, among others, all in a strict relationship. Thus, determining and consolidating all these magnitudes in a single equation to work with becomes very complicated, besides being a nonlinear system. The present study seeks to build an Ant Colony Optimization (ACO) algorithm for identifying the transfer function of a ball-plate system based on the real response of the system to a controlled stimulus, and once the plant is identified, to calculate the constants of a Proportional Integral Derivative Controller (PID) that will allow a better performance of the system, thus to observe the adaptability of this algorithm to solve different problems. As test equipment, a ball-plate system model 33–240 manufactured by Feedback Ltd. is used, from which real data are obtained to be used by MATLAB/SIMULINK software, in which the optimization algorithm and control loops are developed to observe the response obtained from both the identified plant and the designed controller. Therefore, this research presents a real approach to this system where the use of a detailed ACO algorithm reviewed through simulation and real work is an important and effective solution to this problem, while would become too costly to solve with a traditional method.