close
close

Robots learn Greifen vom Menschen

Robots learn Greifen vom Menschen

Learning machines
Robots learn Greifen vom Menschen




Quelle: Innovationscampus Mobilität der Zukunft, Universität Stuttgart

4 min Lesedauer

Anbieter zum Thema

The IFL of the Karlsruher Institute of Technology and the IAS of the University of Stuttgart are studying a robot, male training through apprenticeship. Dafür haben sie mit dem ICM-Zukunftslabor Hapt-X-Deep un in Deutschland einzigartige Forschungsinfrastruktur aufgebaut.

Teamarbeit for die Zukunft: Edgar Welte and Juniorprofessorin Rania Rayyes in the Zukunftslabor Hapt-X-Deep – they train with an experienced robot system, the men train in training.

(Photo: Amadeus Bramsiepe, KIT)

Edgar Welte schnappt put his finger on nach der Luft and brought den Roboter dem Menschsein a handgriff näher. So, this is simply the most important in naher Zukunft sein. Today, the Explorer uses Hapt-X’s virtual reality hand for the robot and the humanoid fern Shadow Dexterous Hand, which must be directed to a work shop.

He was trained by doctor and junior professor Rania Rayyes from the ICM Zukunfts Hapt-X-Deep laboratory at the Institute of Technology and Logistics. System (IFL) of the KIT, a Robotergreifsystem zu entwickeln, das menschliche Tätigkeiten durch Imitation erlernt. Germany first worked with the complete system from the company Shadow Robot in Karlsruhe, and the research study took place within the Institute for Automation Technology and Software (IAS) of the University of Stuttgart.

The Institute takes care of an experienced robot with a Greifsystem, der Schnell and flexible, also zuverlässig and safe auf veränderte Anforderungen, Productdesigns or Materielalien reactieren kann. The funding amounts to 200,000 euros for the ICM infrastructure.

Learning by imitation: learning robots as auszubildende

Programming robots for industrial tasks is a longer process. When the Codes follow the Tests and the Neuschreiben, so often the Machine uses the Process to be corrected. This learning phase is related to the production technology of production, and a certain approach from top management is occurring. “A robot takes care of it, there is no more time in the project, and a new mitarbeiter einzulernen,” said Edgar Welte.

Präzise Nachahmung: The humanoid Shadow Dexterous Hand in the Hapt-X-Deep development work imitates the practice of a male hand – a complete blueprint for the intuitive knowledge of robots.

(Photo: Amadeus Bramsiepe, KIT)

This Ziel zu erreichen will be the Team den Robotern das Lernen von Menschen beibringen. Wie Auszubildende von ihren Meisterinnen et Meistern lernen, soll ihr Robotergreifsystem einst von un Operator lernen, wie sie sie neue Werkzeuge nutzen, unterschiedliche Materialien anfassen, ganze Arbeitsschritte ausführen oder auf Veränderungen in Produktionsprozessen reacts. “We offer you an autonomous system for imitation learning and deep reinforcement learning. The robot learns the methods through interaction with humans, intuitively and easily,” explains Rania Rayyes. Durch den Ansatz verkürzen sich neben dem ursprüngliche Programmierprozess auch Umrüstzeiten.

Hapt-X-Deep three Mensch-Machine-Communikation voran

Menschliches Lernen ist ein Informationsaustausch. Die Beteiligten miteinander, performeren visual Informationen et stellen Rückfragen. I am Zukunftslabor Hapt-X-Deep communicating searches with the robot by date. Edgar Welte generated his own design with Hapt-X-Glove data and created the Cobot carrying bags, as well as its clothing and tools or tool packaging. Thanks to the 20 racing engines, the Shadow’s dextor hand can use hand working tools quickly and quickly, better than other weapons on the market. Pulse sensors in the mechanical fingers respond to direct feedback from the operator. Edgar Welte spürt durch petit Luftpolster im Hapt-X-Glove, sobald das Gewicht zunimmt oder the Werkzeug faux in der Hand liegt. Then faster and softer and the Greifer is there for me.

Stay tuned for Automation App Price 2024!

automation app price 2024

    

Do you have an application or software in use, are the vertreterns from the automation industry suitable for your work? Then you can buy and download them until December 4, with the price of the automation application, the best automation applications auszeichnet.

    

Please, please!

Der Forscher korrigiert den Fehler des Roboters in Echtzeit. Also with a different function, the machine itself is operated by the Künstliche Intelligenz self-supporting. “Les Korrekturen unmittelbaren können wir die Fähigkeiten nos Systems schnell erweitern et ersparen uns Monate an Arbeit für die Neuprogrammierungen,” writes Rania Rayyes.

About innovationsCampus Mobilität der Zukunft

Mobility and production of the Zukunft are efficient, effective and common in Baden-Württemberg. We offer new technologies – innovative technologies and innovative products. The Ziel des Innovationscampus Mobilität der Zukunft (ICM) is here to guide you. The ICM brings together the University of Stuttgart and the Karlsruher Institute of Technology (KIT) for research and innovation competence, a comprehensive package and flexible new technology for business, new analysis for testing and the basis for spring innovation for society. ICM is one of the company’s largest mobility and production initiatives in Germany.

The Zukunftslabor Hapt-X-Deep has a dual function. It is the hardware for research and integration of software management technologies and the demonstration model for high functionality. “It is for me a unique collaboration that I am for my research work with such a high-level work that the work can do,” said Edgar Welte. Mit seiner Promotion zum Thema “Interactive Imitation Learning for Dexterous Manipulation” is and has been included in the publication of the Gesamtsystems liefern. Daneben is also available in Hapt-X-Deep for the Sensorik project, an alternative teleoperation method to managing motion tracking or regulating the underlying finger.

Das schließt auch Cooperationen mit Unternehmen wie Bosch AI oreren Hochschulinstituten en. “Hapt-X-Deep is the most efficient solution for research,” writes Rania Rayyes. One Day was created by the IAS of the University of Stuttgart with the group of junior professor Andrey Morozov. The cooperation projects target the perspectives of Stuttgart methods, safety and security of assistive robots for leisure.

Robot optimization is self-sustaining during operation

Digital learning programs are used as systems testing tools for a deep learning system, the deep learning system helps the robot to use anomalies and use Fehler. This is the basis for errors that can be performed independently. We developed this software project in Stuttgart and tested it in the IFL Reallabor at KIT. Hapt-X-Deep is a modern state-of-the-art infrastructure and structure that draws inspiration from the ICM to bring the best knowledge from the Landes, a means of public transport for learning human robots.

(ID:50259995)