Integración de un sistema de reconocimiento de gestos de la mano basado en el sensor Myo Armband con el simulador de robots Coppeliasim para el control de un manipulador virtual de 6 grados de libertad.

Abstract

This work presents the design, simulation and performance comparison of two control techniques: a PID controller based on minimum norm and a controller based on linear algebra, both for a virtual UR5 robotic manipulator. These controllers are integrated with the hand gesture recognition system developed in the PIGR-19-07 Research Project of the Escuela Politecnica Nacional. This system is capable of recognizing five different hand gestures using electromyographic signals acquired by a Myo Armband sensor, which is placed on the user's forearm. The virtual environment, where the simulation tests of the robot manipulator controllers are carried, is developed in the robotic simulation software CoppeliaSim. The controllers are designed based on the kinematic model of the robot, so the development of this model is presented starting from its basic geometric relationships. The performance of each controller is also compared through the Integral Absolute Error (IAE) criterion from functional tests with and without the gesture recognition system. In addition, the implementation of a graphical interface developed in Matlab's App Designer is detailed, which allows the user to interact with the robot simulation, to choose between its different operating modes, and to monitor its status.