idw - Informationsdienst
Wissenschaft
On behalf of German Edge Cloud (GEC) and the Innovation Cluster 5G Berlin e.V., the Fraunhofer Heinrich Hertz Institute (HHI), together with the Fraunhofer Institutes FOKUS and IPK, has put into operation a real-time capable communication infrastructure for testing industrial applications in 5G standalone campus networks. As one of the first operational distributed 5G standalone campus networks in Germany with heterogeneous radio technology, it is intended to serve as a test environment for future real-time capable edge cloud applications in Industry 4.0.
Germany is one of the international pioneers in the provision of 5G for private use and the establishment of so-called campus networks, i.e. local and customized 5G mobile communications networks. The 5G communications infrastructure now in operation is intended to further extend this lead. The distributed campus networks use fiber optic connections to link an edge cloud at Fraunhofer HHI with Fraunhofer FOKUS and with an extensive machine park as well as 5G radio cells at Fraunhofer IPK in real time. The edge cloud acts as a platform for both the 5G core network (Fraunhofer FOKUS' Open5GCore) and the artificial intelligence (AI)-based software for evaluating sensor data (HHI) and controlling the machines (IPK). In addition to industrial 5G applications, research teams will use the infrastructure to further develop network technologies moving towards 6G, especially with regard to a future open 5G/6G ecosystem.
All three Berlin sites feature the IEEE Time Sensitive Networking (TSN) standard for wired real-time communication as well as 5G campus networks. This setup of networked and distributed network infrastructure including machinery offers unique conditions for innovative, realistic proof-of-concept demonstrations. The communication infrastructure is now available to customers and partners from industry and research for joint projects.
The researchers have already realized first demonstrations. For example, to test the industrial application of remote video inspection for quality control, a production cell at Fraunhofer IPK consisting of a robot, a conveyor belt and flexible 2D industrial cameras was controlled from Fraunhofer HHI in real time via TSN over fiber optics together with the industrial partner GEC. The machine vision models developed by Fraunhofer HHI for capturing hand positions and hand gestures of the users, component identification and position tracking processed the video signals from the remote production cell in real time on the edge cloud. Communication between the sites was accomplished using optical wavelength division multiplexing (DWDM: Dense Wavelength Division Multiplexing). Thereby, the researchers were able to achieve extremely low latencies for the transmission of the video signals from the machine hall to the edge cloud as well as for the control signals in the opposite direction. This is one of the first TSN connections via DWDM that has been implemented in an industrial automation scenario to date.
"In this project, we were able to implement an ambitious video analysis demonstrator in which our AI-based component and hand gesture recognition was outsourced to the edge cloud with the IPK's actuator control and networked via TSN and 5G. This provides an important basis for making future human-robot interaction scenarios more flexible while reducing costs," says Paul Chojecki, project manager at Fraunhofer HHI for the video analysis use case.
In addition, an AI service platform for privacy-preserving machine learning (ML) at distributed locations was developed at Fraunhofer HHI in cooperation with GEC. This AI service platform (DLFi - Distributed Learning Framework) enables the collaborative training of ML models through distributed learning without the need to transmit the training data to a central location. By doing so, it significantly contributes to the security of sensitive data in an industrial context.
Another established use case is cloud-based control and navigation for automated guided vehicles (AGVs). Together with industry partner Gestalt Robotics, the team has developed control and navigation for a 5G-enabled AGV designed to run in the edge cloud. In addition, they were able to virtualize the control of a robotic cell in the form of micro-services and demonstrate the collaboration between AGV and robotic cell in the production process. In this process, the stable communication link with low latency between AGV, robot cell and edge cloud proved to be an important prerequisite.
Beyond the industrial application, the open campus network will advance the research of network technologies leading to 6G, especially with regard to a future open 5G/6G ecosystem. The researchers are exploring this based on end-to-end virtualization of the infrastructure, disaggregation of radio access network (RAN) and core network functions, and the establishment of open interfaces as proposed in current OpenRAN approaches.
"With this comprehensive infrastructure, we and the participating partners have created a unique test environment for new industrial applications, communication and cloud technologies in an urban environment," comments Dr. Johannes Fischer, group leader at Fraunhofer HHI and overall project manager. "It offers our partners a wide range of opportunities to test new prototypes and functional models in a relevant operational environment."
"With the infrastructure consisting of a 5G network and edge cloud now available to us, we at Fraunhofer IPK have the opportunity to determine the requirements for automation technology in an industry-oriented environment. This allows us to pursue targeted development. Based on the experience gained, we can provide our partners from industry with the best possible advice on the use of similar infrastructure," says Prof. Dr.-Ing. Jörg Krüger, head of the Automation Technology division at Fraunhofer IPK.
"The vendor neutrality of our distributed Fraunhofer campus network promotes an innovative, open network infrastructure, as it allows the most powerful components from different manufacturers to be combined - just as needed," says Prof. Dr. Thomas Magedanz, head of the Software-based Networks business unit at Fraunhofer FOKUS. "I am firmly convinced that the trend toward highly specialized private (campus) networks will shape the future of mobile telecommunications. Virtual, i.e. software-based, networks are of central importance here. They enable efficient realization of application-specific and, in particular, dynamically optimizable campus networks and, with regard to 6G, continuous improvement of the network without having to replace expensive hardware."
"The Innovation Cluster 5G Berlin has contributed to the implementation of the innovative Industrie 4.0 application scenarios with its open and experimental test environment for 5G standalone campus networks," comments Prof. Dr. Erich Zielinski, board member at Innovationscluster 5G Berlin e.V.
The communication infrastructure marks the conclusion of the research project "Establishing a demonstrator for the use of 5G, AI and AGVs in industrial scenarios", which the Fraunhofer institutes jointly conducted on behalf of GEC and 5G Berlin e. V.
https://www.hhi.fraunhofer.de/en/departments/pn/products-and-solutions/dlfi-the-...
5G Edge Cloud
metamorworks/Shutterstock.com
Merkmale dieser Pressemitteilung:
Journalisten, Wissenschaftler
Elektrotechnik, Informationstechnik, Maschinenbau, Medien- und Kommunikationswissenschaften, Wirtschaft
überregional
Forschungsprojekte, Kooperationen
Englisch
Sie können Suchbegriffe mit und, oder und / oder nicht verknüpfen, z. B. Philo nicht logie.
Verknüpfungen können Sie mit Klammern voneinander trennen, z. B. (Philo nicht logie) oder (Psycho und logie).
Zusammenhängende Worte werden als Wortgruppe gesucht, wenn Sie sie in Anführungsstriche setzen, z. B. „Bundesrepublik Deutschland“.
Die Erweiterte Suche können Sie auch nutzen, ohne Suchbegriffe einzugeben. Sie orientiert sich dann an den Kriterien, die Sie ausgewählt haben (z. B. nach dem Land oder dem Sachgebiet).
Haben Sie in einer Kategorie kein Kriterium ausgewählt, wird die gesamte Kategorie durchsucht (z.B. alle Sachgebiete oder alle Länder).