The team Agents investigates these concepts in real-world (commercial) applications as well as in virtual environments. It concentrates on agent-based modelling and software development as well as on specific technologies for the design of agent behaviour. More recently a stronger focus was set on motion modelling and synthesis for agent design in 3D simulations. In this context learning techniques with statistical models and neuronal networks (deep learning) are investigated.
- Agent Architectures
- Model Driven Development
- Communication-Based Optimization
- Modelling of Agent Behaviour with Focus on Motion Modelling and Synthesis
- Statistical Motion Models
- Physics-Based Motion Models
- Automated Learning of Motion Models
Agent-based systems have been developed for different application domains, mainly as prototype systems. However, for Saarstahl AG a shop floor control system for steel production has been developed which is used 24/7 in the steelworks of Saarstahl AG in Völklingen. Currently the following applications are investigated:
- Worker simulation for 3D simulations of assembly processes.
- Pedestrian simulation for the simulation of trafic scenarios to generate synthetic data for the validation of autonomous cars.
- Architectures for Industrie 4.0 simulations
- Optimization of logistics processes
- REACT: The overall goal of REACT is a systematic, safe and validatable approach to developing, training and use of digital reality with the goal to ensure safe and reliable acting autonomous systems – especially in critical situations.
- Hybride Teams in wandlungsfähigen, cyber-physischen Produktionsumgebungen (Hybr-iT): The goal of Hybr-iT is the design and industrial evaluation of hybrid teams in which human users and robots with the help of software-based assisting systems collaborate in an smart environment in the setting of a industrial production unit. To this end a unifying view will be applied to human robot cooperation, intelligent planning envnironments, assisting sysstems, and knowledge based rototics.
- prospective.HARVEST: investigtes and develops a prototype for the optimization of the logistics process for corn harvesting. The implementation will be based on a architecture of complementary services.
- BEinCPPS: BEinCPPS Innovation Action aims to integrate and experiment a CPS-oriented Future Internet-based machinefactory-cloud service platform firstly intensively in five selected Smart Specialization Strategy Vanguard regions
(Lombardia in Italy, Euskadi in Spain, Baden Wuertemberg in Germany, Norte in Portugal, Rhone Alpes in France),
afterwards extensively in all European regions, by involving local competence centers and manufacturing SMEs.
- Jumyung Um, Klaus Fischer, Torsten Spieldenner, Dennis Kolberg: Development a modular factory with modular software components. Procedia Manufacturing, 11:922-930, 2017.
- Erik Herrmann, Martin Manns, Han Du,Somayeh Hosseini, and Klaus Fischer: Accelerating statistical human motion synthesis using space partitioning data structures. Computer Animation and Virtual Worlds, 28:3-4, 2017.
- Stephan Busemann, Jörg Steffen, and Erik Herrmann: Interactive planning of manual assembly operations: From language to motion. Procedia CIRP, 41:224–229, Elsevier, 2016.
- Han Du, Martin Manns, Erik Herrmann, and Klaus Fischer: Joint Angle Data Representation for Data Driven Human Motion Synthesis. In: Roberto Teti (Eds.): Proceedings of the 48th CIRP Conference on Manufacturing Systems. CIRP Conference on Manufactoring Systems (CIRP CMS-2015), 41:746-751, June 24-26, Ischia (Naples), Italy Elsevier, 2016.
- Han Du, Somayeh Hosseini, Martin Manns, Erik Herrmann, and Klaus Fischer: Scaled functional principal component analysis for human motion synthesis. In: Proceedings of the 9th International Conference on Motion in Games. ACM SIGGRAPH Conference on Motion in Games (MIG-16), 9th International Conference on Motion in Games, October 10-12, Burlingame, California, USA, Pages 139-144, ACM, 2016.
Team and Contact
- Dr. Klaus Fischer (Head)
- Erik Herrmann
- Han Du
- Somayeh Hosseini