Development and implementation of a natural interface to control an industrial hydraulic robot arm

Abstract

This paper discusses the architecture, the development and the implementation of a natural interface based on a vision system to control an industrial hydraulic robot arm with three Degrees Of Freedom (3 DOF). The project consists of using the natural movements of a human operator to capture commands to move the robot arm. By moving his upper limbs, hands and arms, the human operator generates control signals through an architecture which is based on a Microsoft Kinect for Windows and an electronic system developed with a microcontroller to trigger the proportional control valves of the hydraulic robot arm. From the architecture we show that we control industrial robotics systems using only the visual natural movement of human body without any other sensing device. We implemented a prototype and gathered experimental results that helped us validate our approach and architecture. The system was built to meet industrial accuracy while respecting safety standards.