Human-robot interface for remote control via IoT communication using deep learning techniques for motion recognition

  • Dieisson Martinelli UTFPR
  • Jonathan Cerbaro UTFPR
  • João Alberto Fabro UTFPR
  • Andre de Oliveira UTFPR
  • Marco Antonio Teixeira UTFPR


Even though there are several recent advances in robotics systems, making them more autonomous, there are still many activities that still require a human operator to continuously control the robot’s behavior. The main problem with this is that many interfaces require some specific and fixed control actions by operators, hindering their mobility and even putting their lives at risk in extreme situations. Some studies have already proposed solutions to this problem using computer vision techniques, eliminating the need for the operator to manipulate specific equipment and be in specific positions to control the robot. However, this approach does not eliminate the need for the operator to be physically present at the operation site, due to the type of connection between the equipment, limiting the connection to just one network. This article aims to solve this problem using the MQTT protocol to develop an IoT communication structure between machines on different networks running ROS. This strategy has been promising, eliminating the need for machines that use ROS to be on the same network to communicate with the same efficiency as a local connection. This form of implementation has some advantages in its usability, such as the safety and convenience of the operator who can perform the tasks remotely and from any machine, the safety of the equipment, since heavy processing machines can be separated from the robot that will be put at risk during operation, logistics, having only the need to transport the robot to the task site while the operator and the processing machine can be in another location and with this the lowest maintenance cost, as the only machine and equipment in constant danger will be the robot.
Palavras-chave: Robots, Servers, Task analysis, Cameras, Robot vision systems, Robot kinematics, Protocols, IoT, MQTT, ROS, OpenPose, Teleoperated, Human-Robot Interface
Como Citar

Selecione um Formato
MARTINELLI, Dieisson; CERBARO, Jonathan; FABRO, João Alberto; DE OLIVEIRA, Andre; TEIXEIRA, Marco Antonio. Human-robot interface for remote control via IoT communication using deep learning techniques for motion recognition. In: SIMPÓSIO BRASILEIRO DE ROBÓTICA E SIMPÓSIO LATINO AMERICANO DE ROBÓTICA (SBR/LARS), 17. , 2020, Natal. Anais [...]. Porto Alegre: Sociedade Brasileira de Computação, 2020 . p. 31-36.