Service robotics
What is behind the SeRoNet project
Developing robot solutions and, in particular, service robots more efficiently and bringing them to market - this is the goal behind the online brokerage platform robot.one, which is being developed as part of the joint project SeRoNet.
It is hard to imagine large-scale production without robots. The situation is different for smaller quantities in small and medium-sized companies or in the skilled trades, where robot-assisted automation has so far mostly been dispensed with. Although the vision of Industry 4.0 - individualized, networked and customer-specific production of small batch sizes - needs to be successfully implemented here too, this requires solutions with manageable costs, programming and embedding efforts. Traditional industrial robots have therefore rarely been economically viable for SMEs.
Service robots, i.e. robots outside of fully automated environments, are fundamentally better suited from an economic and cost perspective. Although great potential is seen for these in both the private and professional sectors, the number of applications in many industries has fallen short of expectations. Why is that?
The biggest challenge at the moment is the configuration and programming of service robots. In particular, it is important to make software components easily reusable within a specific robot solution in order to save costs during system integration. Today, this usually accounts for 60 to 80 % of the investment in service robots. This is where IT platforms come into play: in future, they should enable a greater division of labor in the development of systems. In the near future, ideas based on existing and reusable solutions, available standard components and service offerings could be implemented more quickly and cost-effectively, making them available to SMEs.
With this in mind, five research institutes, a system integrator, three end users and a robot manufacturer have joined forces in the SeRoNet project and jointly developed the 'robot one' platform.
Functionalities of the robot one platform
Its most important aspect and unique selling point is its mediation functionality. The platform is ultimately intended to connect component manufacturers, system integrators and end customers. Based on problem descriptions to be formulated, it supports end users in finding experienced system integrators and suitable hardware and software components. Similarly, system integrators can submit suitable offers to potential customers and select suitable components. Finally, component suppliers can use the platform to position their products in line with demand.
Robot one controls the flow of information in this regard, which makes the platform itself part of the value creation network. An important criterion for the description of requirements and solutions is to protect this data appropriately against unauthorized access so that the data owner can control what happens to their information. In many cases, abstract framework conditions are also sufficient as a description of the problem or solution. For example, these can contain details such as the application sector, environmental conditions for operation (brightness, temperature, etc.), workpiece properties (material, moving masses, etc.) or accuracy requirements. In addition, solution descriptions can reference the components used and the associated problem. All data is used within the platform to support the search for providers or partners when new problem descriptions are entered.
Ultimately, this approach promises various advantages for the different players: component manufacturers can market their products and technologies more effectively. The ability to place products in the SeRoNet catalog could open up previously unreached markets. For system integrators, the platform provides a transparent view of existing components - whether as open source or proprietary software components or as 'intelligent' hardware modules. This increases the choice of possible system modules and enables previously unknown components to be integrated. In addition, suitable components for an ordered service robot system can be found more quickly. Finally, end users benefit from the cost savings resulting from more efficient system integration and reusable modules. On top of this, the robot one component catalog also gives users a quick overview of possible partial solutions and technical developments on the supplier side.
Modeling with Industry 4.0 tools
The technical basis of this collaborative work is a uniform modeling of the components used to implement robot systems. This makes them easy to combine and find. Part of the models is, for example, the description of the communication behavior, including whether a component publishes data cyclically (pub/sub model) or uses service calls (RPC) for communication.
However, these technical models are not sufficient to make the components discoverable by robot one on the basis of their (technical) properties. Consideration of a typical life cycle of an industrial plant shows that the application-relevant, functional properties of components play the overriding role, particularly in the planning phase. When selecting a distance sensor, for example, the decisive properties are the measuring range or the measuring accuracy. Only when suitable sensors have been found for these parameters can the selection be narrowed down by other criteria such as purchase price, service life or interfaces.
SeRoNet demonstrator: The picking system will enable sensor-controlled, adaptive goods removal.
© FZI KarlsruheIn addition to the technical information, virtually every component also includes other general or domain-specific information such as existing service contracts, payment models or warranty periods. The total amount of this information can be extensive and heterogeneous. The standardized representation of such information in a machine-processable form is an important task in the course of Industry 4.0 and in SeRoNet. It is made possible by the so-called 'Administration Shell' of the 'Plattform Industrie 4.0' initiative.
A key result of this initiative is the 'Administration Shell in Detail' (VWSiD) specification published at the end of 2018. This defines a metamodel of the asset administration shell, the digital representation of every Industrie 4.0-compatible asset - for example, components of a plant, entire production systems, but also workpieces - as well as some mappings of this model to specific technologies (currently XML and JSON). Other mappings - such as RDF or OPC UA - are being considered or are in preparation. However, it is not the task of the VWSiD specification to define the specific content of asset administration shells. Due to their diversity, this is the task of dedicated organizations/committees.
Modelling and standardization
In modelling, SeRoNet takes up the concepts of the asset administration shell and thus carries out preliminary work for further standardization work, validates the previous standardization results in practice and thus ensures the future viability and applicability of the project results in the industry. In this way, the standardized component modeling - in particular the technical data sheet - can be brought into the VWSiD form. At the same time, further description models can be defined for hardware or software components frequently used in service robotics.
The central goal of the robot one platform, namely to find the most suitable components for a task or to bring together suitable actors, will lead to a comprehensive database in the medium term, in which the descriptions of the components are preferably stored in the form of administration shells. RDF, the 'Resource Description Framework' of the WorldWideWeb Consortium, is used within the mediation platform for this purpose. A method for mapping the VWSiD metamodel to RDF is currently being developed in SeRoNet in close cooperation with the Industry 4.0 committees and will be evaluated in the project.
As OPC UA defines a powerful information metamodel as a recognized industry standard, it is to be expected that the administration shells or at least submodels of many assets will be described in OPC UA in the future. This applies in particular to assets that are also OPC UA components, i.e. that meet the OPC UA Foundation's requirements for interfaces and information models. Components developed with SeRoNet are automatically such OPC UA components. By using the standardized VWSiD metamodel, it will be possible for robot one to import the component descriptions provided in OPC UA format with little effort.
In order to map the heterogeneity of all potential assets, the VWSiD metamodel offers many degrees of freedom in modeling. Even if this flexibility is necessary in order to map all conceivable scenarios, restrictions can be defined for specific problems in order to increase interoperability. For particularly important and generally valid use cases, such restrictions can be formalized in the form of additional standards or profiles. In SeRoNet, for example, proposals for mapping hierarchies of component classes are being developed and implemented.
Semantic backend
The robot-one platform uses a semantic backend for structured and expandable data storage. This database based on RDF organizes the assets described by the Industry 4.0 asset administration shell and other technical models into user-oriented classes and role-specific views for the platform. Existing software and hardware components, central roles for robot one (customer, system integrator and component manufacturer) and the resulting connections (e.g. experience with components from system integrators) are stored in a structured graph.
To ensure that the modeled instances contain all the necessary information and comply with existing conditions (e.g. a link to component documentation), newly inserted or modified components are automatically checked against a model schema. If the schemas are not fulfilled, the exact sources of error can be found and shown to the creator of the component. To enable access by platform users, the database is prepared for a faceted and full-text search. Furthermore, robot one provides forms for editing the data by their respective owners.
In order to generate an initial data set for robot one, the project offers several opportunities to apply for funding for the integration of existing software and hardware components. Participants receive early access to the tools from SeRoNet and can be the first to be represented on the platform with their products.
Call for sponsored participation
Participants are also invited to help shape the design of the models and the platform itself. The first call is expected to be published on September 2, 2019 and is aimed at providers of mechatronic or pure software robot components. The components to be submitted should be reusable, have a SeRoNet-compliant information model and communicate via the specified backend - i.e. OPC UA. The funding of 20,000 to 40,000 euros is intended to finance the additional work involved in adapting the components to the SeRoNet models and OPC UA interfaces. Funds of 1 million euros are available for all calls for participation combined.
Three rounds of calls are planned, each a few weeks apart, to deliver a wide range of robot components for robot one. Examples of components include sensors and actuators for detecting the environment or interaction as well as software components that enable collision-free navigation with the aid of sensor data, for example. The project will also integrate components from interested providers into the platform after a quality assurance process.
Authors:
Dr. Björn Kahl is project manager at the Fraunhofer Institute for Manufacturing Engineering and Automation IPA;
Yevgen Kogan is team leader at Kuka Germany;
Jörn Steffen Menzefricke is a research associate at the Heinz Nixdorf Institute, University of Paderborn.
You might also be interested in
IFR forecasts growth pause at record level
The 'World Robotics Report 2019' published by the International Federation of Robotics (IFR) reports a global record sales value of 16.5 billion US dollars for industrial robotics in 2018. After a pause in growth this year, double-digit growth is...
read more...Robotics Institute Germany co-founded
Fraunhofer IPA is a co-founder of the Robotics Institute Germany, which aims to bundle Germany's cutting-edge research into AI-based robotics.
read more..."Manufacturing-X data room preliminary study" presented
Manufacturing-X is seen as the next step on the road to Industry 4.0, but how could a data room be designed for the data ecosystem? A study presented by VDMA and ZVEI gets to the bottom of this question.
read more...How companies use data for product development
Where do companies stand when it comes to collecting data from products and using it to develop new products? Researchers from the Fraunhofer Institute for Manufacturing Engineering and Automation IPA have investigated this in a study.
read more...Fraunhofer project determines properties of coatings virtually
In the "PaintVisco" project, researchers at Fraunhofer IPA are modeling the development and processing of paints. The bat was the model for the development of a new simulation program that can be used to virtually determine the exact properties of...
read more...The digital value stream analysis
A research team from Fraunhofer IPA is digitizing value stream analysis together with the Stuttgart-based software provider Ifakt. This should enable this optimization method to be carried out in the future with significantly less effort and almost...
read more...The top articles in January 2022
Obstacles and opportunities characterize the top 10 most-read articles in January. Click through the obstacles in companies and find out what opportunities the year 2022 has in store.
read more...Intelligent sensor detects leaks
To simplify the search for leaks in compressed air systems, a researcher from Fraunhofer IPA is working with Sick to develop an additional leakage service for an intelligent flow sensor. Self-learning algorithms evaluate the measurement data.
read more...Improved safety for service robots
In the S³ research project, researchers from Fraunhofer IPA worked with project partners to develop safety technologies and recognition algorithms for service robots. They were tested on the Care-O-bot 4 service robot in a nursing home application.
read more...