Robotized application based on deep learning and Internet of Things

摘要

This paper presents a way to integrate an industrial robot into Internet of Things and to use it with a deep learning application. Besides of manufacturer's restrictions, which usually exist in an industrial scenario, an easy method to extend and merge the sensorial and decisional systems for robots will be required as part of Industry 4.0. Related to this, the proposed method couples IBM Watson IoT cloud-based platform, a Node-RED cloud application, a deep learning mechanism with TensorFlow (this being applied for a computer vision case study), and an old generation industrial robot. Several conclusions highlight the tradeoff of using IoT and deep learning solutions for a real manufacturing environment.