Collaborative work with Stéphane Thiery, Éric Nyiri and Olivier Gibaru

This work led to a publication as a conference article at ICRCA 2021.

In recent years, industrial robots have become more collaborative thanks to better sensors and higher level programming libraries. Yet, in real world scenarios, flexibility and interaction abilities of robots remains far from the natural interaction between two human co-workers. Thus, we are developing a smart robot agent (SRA) that can incrementally learn with a human in a teacher/learner setting called Interactive Task Learning (ITL).

This video presents a high level overview and first prototype of this architecture for Industry 4.0. It is validated on a real Universal robot (UR10e) : an operator teaches the robot the task “give object”.

Video main sections :

0:00 Robots as smart assistants for Industry 4.0

1:29 Why a robotic architecture ?

2:12 Towards an Interactive Task Learner

3:15 High level overview of the architecture

4:00 Representations choices through an example

9:00 Validation on a real robotic system

10:40 Perspectives