A-DRZ – Aufbau des deutschen Rettungsrobotik-Zentrums
Dr. Constantin Houy, Peter Pfeiffer
Abstract: The project A-DRZ aims at setting up the German Rescue Robotics Center (DRZ) (https://rettungsrobotik.de). Over the next few years, this competence center for rescue robotics will be established in Dortmund, where mobile robot systems for civil emergency response will be researched and developed in a living lab. Currently, robots are typically used as fully tele-operated tools in emergency missions. However, with increasing autonomy, robot systems are supposed to become agents that actively participate in a mission. For this they need an understanding of the emergency mission objectives, the tasks of the human-robot team and their execution. Only in this way robot systems can perform tasks as expected and take initiative in a meaningful way.
Using techniques of speech processing and process modeling, DFKI (participating with the three research departments MLT, IWi and EdTech) explores and models the necessary cooperation and communication processes and develops technologies that enable robots to understand and follow emergency missions. We develop innovative solutions that use process analysis techniques and knowledge services to support efficient and effective mission execution through the resulting understanding.
The project is funded by the German Ministry of Education and Research (BMBF).
AdjUST – Automatisierung in der Konfiguration von Unternehmensinformationssystemen der Textilbranche über Methoden der KI und Referenzmodellierung
Philip Hake, Brian Willems
Abstract: In the course of digitalization and globalization, companies in the textile company are faced with the challenge of professionalizing their processes. This concerns the core processes in administration and production as well as, for example, the processes in sales. As smaller companies in particular only have small budgets, individual developments are usually ruled out, and so is the adaption of expensive standard software. In order to meet these special requirements, innovative techniques of process analysis and artificial intelligence are used in the described project to enable companies to use software solutions for process handling in the most resource-efficient way possible. The execution data generated from system use is used by the system operator to successively develop inductive reference models for the textile industry, from which other companies then can benefit. At the same time, AI technologies are used to enable automated, intelligent customization of the system for new customers. The goal is to provide companies with an ERP system that is easy to use. By making (anonymized) usage data available, best practices can be disseminated within the industry and thus processes that do not primarily add value can be digitized cost-effectively.
AKKORD – Vernetzte und integrierte Anwendung industrieller Datenanalyse für die wertschaffende, kompetenzorientierte Kollaboration in dynamischen Wertschöpfungsnetzwerken
Steffen Schuhmann, Sabine Klein
Abstract: Industrial data analyses makes innovative possibilities for sustainable optimisation of products and processes for production companies available and makes it possible to initiate new business models and collaborations in value networks. However, current efforts are influenced by the growing realisation that SMEs are only partially able to make meaningful and targeted use of modern analysis technologies on their own. There is a lack of the necessary competencies and implementation strategies in the companies themselves, as well as a lack of strategically aligned service and technology offerings to be able to sustainably utilise the extensive potentials.
The goal of the AKKORD project (networked and integrated application of industrial data analysis for value-creating, competence-oriented collaboration in dynamic value networks) is to enable companies to profitably apply data analysis – both within the company and acrossvalue networks. For this purpose, an online service platform is being developed that includes solutions for comprehensive data integration as well as easy-to-use, standardized data analysis modules and dashboards, and supports the initiation of new collaborations and business models. In addition, the development and provision of modules for competence analysis and acquisition lays the foundation for targeted competence enhancement of employees in the field of data analysis. The Institute of Information Systems (IWi) at DFKI is researching the potential of data analysis for the cross-product lifecycle adaption of business models in AKKORD. Head of the institute Prof. Peter Loos emphasizes: “Process mining on production and product usage data can be used to draw interesting conclusions about product design or the design of the business model. This can also facilitate coordination mechadeeplnisms regarding pricing in the value chain.”
APPaM – Automated Process Planning and Mining
Abstract: The correct execution of business processes is already being examined in companies nowadays with the help of process mining tools. Based on the execution logs of the IT systems, these tools develop human-readable process models that reflect the current state. For the readability of these models, the real executions are simplified, and atomic events are combined into generic alternatives. As a result, these process models are no longer suitable for intelligent planning or automation. Besides this, forecasting models which predict future events already exist, but the conversion into a concrete recommendation for action is still left to the employees. For effective support, a system is needed that automatically recognizes business processes from the system logs and converts them directly into cost-optimized recommendations for action to support employees in their planning. The goal of the project is therefor to develop methods, algorithms and prototypes of the research areas automated planning and process mining. The solution should enable automated planning and execution of business processes.
To achieve this goal, new algorithms that operate at the atomic event level if IT systems and create process models there will be developed first. These models are to be enriched with action alternatives comparison to know process models. Based on this, domain-specific planning problems that can be optimized in a time-efficient and parallelized manner on graphic cars will be defined. In the next phase, the formerly mentioned components and known process prediction models will be transferred into an integrated solution, which will then be implemented as a prototype. The usability of the system will be put to the test in several practical projects from different industries.The solution will be able to show its benefits especially in time-critical business processes. It is the dynamic character of the planning that enables adjustments even after the individual process executions have started. This means that planning can also consider short-term changes or special events. Examples include production planning in the fashion industry or the deployment planning of emergency services. Both areas require dynamic and situational process planning and were therefore identified as test fields for practical use.
AutoReGen – Entwicklung eines Verfahrens zur automatisierten Überprüfung der Rechtssicherheit und der Generierung rechtssicherer Rechtstexte für Internetseiten auf Basis der Methoden maschinellen Lernens
Andreas Emrich, Michael Frey
The research goals of the DFKI in the project are:
- Text mining of existing legal texts (judgments, laws, legal sources) using Hadoop and Spark.
- Based on the results of text mining, the development and maintenance of a user matrix and a text matrix.
Development of a recommender system based on the calculation of a utility matrix, for the recommendation of a suitable legal text.
COC-TT – Center of Comptence Tax Technology
Prof. Dr. Peter Fettke, Martin Scheid
Abstract: The vision of a tax function of the future – researching innovative artificial intelligence technologies and accompanying the transfer of innovation to design the tax workplace of the future.
The Center of Competence Tax Technology (CoC-TT) involves a targeted thematic collaboration between the German Research Center for Artificial Intelligence (DFKI) and the tax consulting company WTS. It pursues the goal of supporting the transfer of innovative technologies and solution approaches from the field of artificial intelligence for tax applications into practice.
The following three central objectives, among others, are pursued in this context:
- Innovative scouting and potential studies: Through the ongoing investigation of current trends and innovations, a continuous evaluation of potentials of modern AI technologies for the field of tax is achieved.
- Creation of applicable solutions: Based on concrete application scenarios from the practice of the participating partner companies, the identified technology potentials are transferred into practically applicable solutions.
- Setup of a demonstration lab for central AI technologies and example scenarios in the control area supports communication as well as exchange between the partners involved and thus reinforces the innovation idea of the competence center.
INTONATE – Integration manueller Arbeitsprozesse in die Smart Factory
Abstract: Modern production plants, also known as smart factories, rely on intelligent production systems that are coupled for manufacturing. The central element here is the networking of the plants, which makes it possible to exchange information and transfer current process parameters. Here, central process control is being broken down more and more, as a result of which decisions are increasingly being made by the manufacturing plants themselves. To exploit the full potential of these smart manufacturing plants, the executed processes must be mapped digitally, allowing processes to be optimized during their execution to the current workload of the smart factory. Here, manual work in particular poses a major challenge. The difficulty lies in collecting information from the flexible manual workflows and placing it in the process context. Today, wearable sensors can already be used to collect movement and vital data. However, these cannot be used in a targeted manner for process planning and optimization in industrial manufacturing environments because they are significantly more fine-grained than usual events in business processes. Consequently, they would first have to be pre-processed, segmented, and aggregated in order to be used meaningfully in processes.
In the INOTATE project, data from such wearables will be collected during the execution of manual bending processes and evaluated in the process context with the aid of artificial intelligence methods to make them useful in process planning.
KEA-Mod – Kompetenzorientiertes E-Assessment für die grafische Modellierung
Dr. Constantin Houy, Peter Pfeiffer
Abstract: Graphic modulation is an established part of information systems higher education and numerous related study programs. The joint project KEA-Mod aims to develop a digital subject concept that improves the teaching quality of graphical modeling. To this end, tools that have so far been isolated from one another and used locally by the collaborative partners, such as task generators, feedback, and assessment systems, will be combined in a uniform overall system and further developed in an application-oriented manner.The result is an “e-assessment platform” that covers various teaching-learning scenarios such as lectures, exercises, and exams. A central feature of the platform is its transferability: It should be possible to use it at different university locations in Germany. The use of the platform is accompanied by qualitative and qualitative methods of evaluation.
KOSMOX – Entwicklung einer neuartigen lokalen kontrafaktischen Erklärungsmethode und -schnittstelle unter Berücksichtigung kognitiver Modelierungsansätze
Nijat Mehdiyev, Lea Mayer
Abstract: In the field of artificial intelligence (AI), black-box models are superior to conventional methods, but they do not provide any explanation of how their decisions and recommendations for action were made. This lack of comprehensibility leads to reservations about the use of these technologies. In addition, questions regarding the legal security, ethics and customer orientation of the applications arise. This is where explainable artificial intelligence (XAI) comes in, explaining decisions and behaviors of the system to the user. However, current research and development activities in the XAI domain ignore relevant aspects: except for the researcher’s intuition about the quality of an explanation, interactive, human, organizational and economic aspects are often ignored. Properties of explanations and problem definitions are also often not considered. Therefore, there is a great need for an approach to the development and design of explanation models and corresponding interfaces – that should be able to be used as a guide in the future.
The aim of the KOSMOX project is to develop a holistic explanation system that presents the decisions of AI systems in a way that can be understood in retrospect. The aim is to create a system that combines approaches of rule-based and simulation-based explanations. In addition, relevant techniques from the field of complex systems will be used. Furthermore, an explanatory interface will be designed and implemented, which is based on insights from the cognitive science and enables interactive communication between the users and the applied AI techniques. In an interdisciplinary team of researchers and industry experts, approaches from organizational science, human-computer interaction, and behavioral economics are also used for this purpose. By highlighting the added value of human and AI collaboration, the aim is to build trust in AI-based applications. The focus is on transparently presenting and explaining the results of machine learning and other AI techniques from the decision-making process to users. When working on the project, attention is to be paid to a cross-industry and cross-sector solution to ensure broad applicability.
Mittelstand 4.0-Kompetenzzentrum Kaiserslautern – Mittelstand 4.0-Kompetenzzentrum Kaiserslautern
Andreas Emrich, Sabine Klein
Abstract: The Medium-sized Business 4.0 Competence Center Kaiserslautern is part of the nationwide network Medium-sized Business Digital. With Medium-sized Business Digital, the German Federal Ministry for Economic Affairs and Energy has been supporting digitalization in small and medium-sized enterprises and the skilled trades since the end of 2015. There are 26 medium-sized Business 4.0 competence centers nationwide, 18 of them have a regional approach and eight of them have a thematic or sectoral approach. The centers have learning and demonstration factories and help SMEs with practice-relevant expert knowledge. The Medium-sized Business 4.0 Competence Center Kaiserslautern began its work in April 2016 and has since accompanied Rhineland-Palatinate SMEs on their path to digital transformation.
The four partners from practice and science are the technology initiative SmartFactory KL, the German Research Center for Artificial Intelligence with the research area innovative factory systems and the Institute for Information Systems, the Technical University of Kaiserslautern with the Chair for Strategy, Innovation and Cooperation and the Chair for Virtual Product Development, and the Institute for Technology and Work. They combine their expertise in the competence center and cover the key topics of digitalization on an interdisciplinary basis: Automation and networking, people and work 4.0, strategy and innovative business models.
In the last three and a half years, more than 1,100 companies have made use of the free services offered by the Kaiserslautern Competence Center. In particular, the project-related support and on-site assistance have met with great interest among SMEs. Every project topic is selected according to the needs of the companies and covers everything from developing a digitization idea, digital process optimization, data networking and retrofitting to Work 4.0 or digital business models. A special and unique offer of the competence center is the training demonstrator PAUL. It was specially developed for flexible on-site use at SMEs and has a modular structure. It conveys the functionality of Industry 4.0 in an easy-to-understand manner. The center also offers a wide-ranging qualification program consisting of workshops and learning videos. A new learning and action platform to explore key topics of digitalization in greater depth is currently being developed.
The Kaiserslautern Competence Center is currently in its second funding phase and has expanded its range of services to include AI-related services for industrial production since July 2019.
The goal is to sensitize SMEs to the technological and economic potential of AI and to disseminate concrete application examples. The goal is to promote the transfer of AI knowledge to companies and its application on site. Relevant fields of application are assistance systems, smart data analyses, and intelligent products and services.
Andreas Emrich, Alexander Berrang
Abstract: The Project MoveTo4.0 helps to accelerate the business of European SMEs in the manufacturing industry to overcome their business challenges with innovative new technology. Within the programme of EIT Manufacturing MoveTo4.0 supports 50 SMEs from all over Europe to create their individual transformation roadmaps towards Industry 4.0.
PROMISE – Prozessoptimierung durch Kombination von Produktionsprozess- und produktspezifischen Daten
Abstract: In today’s process flows, new technologies are increasingly being used which are generating an ever-growing amount of process and product data. This information is already being fed into various systems and made available to users, but they are not used to a sufficient extent. Process analytics can be further expanded in the context of Industry 4.0 in today’s applications – for example, added value can be generated from the integration of different data formats by using additional information to optimize processes. In the automotive industry, the testing phase of new products is becoming increasingly difficult due to high-cost pressure, increased technology changes and shorter time-to-market – although there are already options for failure analysis and status monitoring in the individual domains, there is a lack of a cross-domain system for complex processes. Occurring irregularities within the product life cycle of a product are detected late, so that only reactive measures can be taken. This can result in minor effects, such as a small amount of scrap, but also in major damages. An example of this would be the installation of a defective component in a vehicle that has been handed over to a customer.
To solve the problem described above, a tool based on artificial intelligence (AI) is to be developed that enables cross-domain status monitoring and inherent process optimization -this is to be done over the entire product lifecycle. Anomaly detection, defect analysis and root cause identification are intended to help limit the production of defective products and minimize potential damage. In order to strengthen the user’s trust in the system, the results are to be presented to the quality engineer in a comprehensible way using methods of explainable AI (XAI).
Andreas Emrich, Oliver Gutermuth
Abstract: In addition to avoiding production errors caused by component failures in machines, a challenge in the manufacturing and process industry is to avoid errors caused by the design of processes and their control. However, the complexity of the processes in this environment does not allow a complete consideration of all influencing factors, so that simplifications are made which imply uncertainties. As a result, the structure and parameters of a system are not fully mapped and the system behavior cannot be predicted exactly. In particular, the quality and defects of the product cannot be directly traced back to specific features of the production process since all relevant process parameters and critical interrelationships are rarely known or recorded.
In principle, machine learning methods could be used for such scenarios to identify dependencies between process outcomes and process variables. However, these methods depend on enough data, which are not available for rare or unknown causes of errors.
In the RACKET project, the problem of detecting rare and unknown faults by combining model-based methods and machine learning methods is addressed theoretically and in representative application scenarios. The goal is to develop a comprehensive methodology for the detection of rare and prior unknown fault events that may occur in a complex factory environment.
SmartMobi – Baukastensystem für mobile Industrie 4.0-Apps
Abstract: The SmartMobi research project addresses the challenges of developing individual mobile applications for industrial scenarios. The research projects aims to develop a modular system for the simple creation of mobile applications (apps). A modular structure enables the coupling of different function modules and the exchange of data via standardized interfaces. The design of the apps is supposed to enable the use of a graphical user interface for the most parts, so that simple applications can be developed with almost no programming knowledge. For the creation of complex apps, a special development platform is provided that already includes module templates for specific functions or connectors for data exchange with machines, sensors, or information systems. This means that the development of individual operational apps can be accelerated and designed with significantly reduced effort.
SmartVigilance – Regulatorische Compliance durch KI-basierte Umfeldüberwachung in der Medizintechnik
Philip Hake, Peter Pfeiffer
Abstract: Medical device manufacturers are subjected to strict regulations regarding the safety of their products. Regulations affect both product approval and the post-market phase of approved medical devices. Companies are obliged to monitor the use and application of their products on the market and to take appropriate measures to eliminate defects or minimize risks. The aim of the project is the prototypical development of technologies and automated procedures for the regulatory required marked monitoring and risk assessment in medical technology. An internet based “SmartViligance” platform will automatically capture, analyze, and report to manufacturers dangerous incidents and product defects that are reported by users to regulators and are publicly available. The project uses artificial intelligence methods and technologies – natural language processing (NLP), machine learning (ML) and data analytics. The platform is used to monito the environment (“vigilance”); it is intended to make market monitoring (“post-market surveillance”) more reliable and to support and relieve companies in the medical technology sector in the process.