+**********************************************************************
*
*
*                          Einladung
*
*
*
*                     Informatik-Oberseminar
*
*
*
+**********************************************************************
 
Zeit:        Donnerstag, 05.12.2024, 13:30-14:30
 
Der öffentliche Vortrag findet hybrid statt:
 
Raum:     Raum 5053.2 (B-IT-Hörsaal)/Informatikzentrum, Ahornstraße 55

Zoom:  https://rwth.zoom-x.de/j/64360669610

Referent:     Herr Mohamed Behery, M. Sc.
                      LuFG Informatik 5 (KBSG)
 

Thema:        Digital Shadows for Industrial Robotic Agents in the World Wide Lab

 
Abstract:

The digital transformation of production has become imperative to increase efficiency, adaptability, and innovation required to face the rapidly changing market requirements, supply chain disruptions, and global competition. In addition to initiatives such as Industry 4.0 and the Industrial Internet of Things (IIoT), the Internet of Production (IoP) aims to establish a world-wide network of research and industrial partners through the World Wide Lab (WWL). Such a network requires standardized communication protocols, which need to remain flexible to accommodate the different standards, use cases, and contexts. To this end, the WWL uses Digital Shadows (DSs) to represent the different entities: parts, products, machines, processes, and companies. DSs are context and purpose-specific representations of entities in the WWL, which enables them to replace the omnipresent Digital Twins by representing different aspects of the denoted entities. The thesis offers a road map for achieving the WWL as an application on top of the IoP.
 
The thesis focuses on industrial robotics processes, which involve teleoperated robots or human-robot collaborative assembly and are not fully integrated in the WWL as they lack a standard DS for task representation. To bridge this gap, the thesis proposes using Behavior Trees (BTs), which offer a flexible, modular, and composable task representation for industrial robotics processes. BTs emerged in the video game industry and were adopted by the roboticists but do not yet fulfil all the requirements to be a standard DS. The presented work offers a plan to fulfill these requirements. 
The thesis introduces a framework for extracting and creating DSs for robot teleoperation tasks in industrial processes. It starts by collecting data to extract discrete actions. After that, it learns a library of BTs that represent the commonly executed behaviors. This allows a user to compose more complex behaviors using the learned BTs. The different modules of the framework are evaluated qualitatively and quantitatively. The thesis also proposes and demonstrates three BT extensions increasing BT flexibility to accommodate more use cases. They allow a BT to handle highly dynamic production environments, where it interacts with heterogeneous agents such as humans or robots with different capabilities or knowledge. 
 
The dissertation concludes with a discussion that evaluates the applicability of the proposed framework and extensions to different industrial scenarios giving recommendations for future work.

 
Es laden ein: die Dozentinnen und Dozenten der Informatik

_______________________________

Leany Maaßen

RWTH Aachen University

Lehrstuhl Informatik 5, LuFG Informatik 5

Prof. Dr. Stefan Decker, Prof. Dr. Matthias Jarke,

Prof. Gerhard Lakemeyer Ph.D., JunProf. Dr. Sandra Geisler

Ahornstrasse 55

D-52074 Aachen

Tel: 0241-80-21509

Fax: 0241-80-22321

E-Mail: maassen@dbis.rwth-aachen.de