Simulación de controladores de seguimiento de trayectoria basados en inteligencia artificial para un robot humanoide NAO : Simulación de un control basado en algoritmos genéticos para el seguimiento de trayectorias de un robot humanoide.

Abstract

This document shows the results of the application of an artificial intelligence technique, based on evolutionary computation, to perform the trajectory tracking control of a humanoid robot. The technique mentioned previously is based on genetic algorithms, which use biological principles and concepts obtained from evolutionary theory to solve optimization problems. For the case of study, it is proposed the development of a controller for trajectory tracking based on Lyapunov, which is designed by using the kinematic model of a mobile robot. Initially, the controller is manually tuned for comparative purposes only. After obtaining the initial controller, the genetic algorithm is applied to calculate optimized values for the controller’s constants. This algorithm performs an offline iterative simulation of the closed-loop process to obtain the best values that minimize an objective function determined by the developer of this work. The obtained results are evaluated by calculating performance indices and the objective function based on them. After the calculation of the optimized gains for the controller, the equivalents of these values are obtained, so that they can be applied to perform the control of the NAO humanoid robot. The operation of this controller, as well as its analysis and comparison with a manually tuned controller, are presented in this document; demonstrating that the developed work offers an improvement in terms of system performance.