ESR 12 – Flexible Robotics
The variety of manufactured products is increasing exponentially. A single factory can produce nowadays between hundreds to thousands of different references in a single year. In this context comes up Industry 4.0 paradigm. Among the main goals that it pursues, zero defects and unitary production are two of the most relevant ones. The aim of this ESR is to explore and develop a new approach towards this two of challenges.
The stated proposal is to provide automation capability to quality control tasks. These tasks will be made by robots equipped with smart machine vision devices that can make the inspection in an adaptive strategy. This approach pursues to develop automated inspection system that can adapt themselves to new references. This will be done by merging deep learning techniques and adaptative control models. Deep learning will provide the capacity to recognize defects and to learn what can make a piece defective or not and apply this knowledge to new incoming references. Adaptive control models will provide the robot’s trajectory for each inspection reference by using the digital information for the manufactured piece and by linking it with the input data from the deep learning module. All of this will be integrated with low inertia robots in order that humans and robots can share workspace.
Fundación Tecnalia Research & Innovation (TECNALIA), Spain, PhD registration with UPV/EHU
TECNALIA –Dr. Estibaliz Garrote