Estudio, desarrollo e integración de equipos auxiliares para el control y/o monitoreo de plataformas robóticas, utilización del sistema de reconocimiento de gestos empleado en el Proyecto de Investigación PIGR-19-07 para el comando de un sistema multi-agente.

Abstract

The research project PIGR-19-07 has developed hand gesture recognition algorithms that are designed by employing the MYO Armband sensor and MATLAB on Windows operating system. To verify its applicability, this project uses this system to command the change of the formation shapes of a robotic multi-agent system. This system is implemented in a virtual environment, so the robotic simulation platform is chosen based on the robotic agents and the way they are commanded. The Gazebo simulator facilitates to launch and work with virtual TurtleBot3 robots as system agents and uses the ROS platform for their command, which is optimized for Linux, so the robotic platform is implemented in a virtual machine. The guidelines for the integration with the hand gesture recognition system are established by taking into consideration the actions to be performed with the commands obtained from both: gestures and the IMU of the MYO sensor. The control is based on the tracking of each agent towards its position, so the ROS Toolbox available in MATLAB-Simulink is used for this implementation. This tool allows the communication with the ROS master node using blocks available for Simulink models. An interface for the execution of the multi-agent scene is developed in the virtual machine and another one in MATLAB for the control of the commanded actions. The carried-out tests are based on verifying the performance of the implemented formation algorithm working together with the hand-gesture recognition system and their results focus on the success of commanding the multi-agent system.