05 Fakultät Informatik, Elektrotechnik und Informationstechnik

Permanent URI for this collectionhttps://elib.uni-stuttgart.de/handle/11682/6

Browse

Filters

2019 - 20206

Settings

Author: search.filters.author.Maschler, BenjaminSubject: search.filters.subject.620

Search Results

Now showing 1 - 6 of 6
  • Thumbnail Image
    ItemOpen Access
    Maschinelles Lernen für intelligente Automatisierungssysteme mit dezentraler Datenhaltung am Anwendungsfall Predictive Maintenance
    (2019) Maschler, Benjamin; Jazdi, Nasser; Weyrich, Michael
    Für eine hohe Ergebnisqualität sind Machine Learning Algorithmen auf eine breite Datenbasis angewiesen. Studien zeigen jedoch, dass viele Unternehmen nicht bereit sind, ihre Daten mit anderen Unternehmen, beispielsweise in Form einer gemeinsamen Daten-Cloud, zu teilen. Ziel sollte es daher sein, effizientes maschinelles Lernen mit einer dezentralen Datenhaltung, die den Verbleib vertraulicher Daten im jeweiligen Ursprungs-Unternehmen ermöglicht, zu ermöglichen. In diesem Artikel wird diesbezüglich ein neuartiges Konzept vorgestellt und hinsichtlich seiner Potentiale für intelligente Automatisierungssysteme am Beispiel des Anwendungsfalls Predictive Maintenance analysiert. Die Umsetzbarkeit des Konzepts unter Nutzung verschiedener bestehender Ansätze wird diskutiert, bevor schließlich auf potentielle Mehrwerte für Anlagenbetreiber sowie -hersteller unter besonderer Berücksichtigung der Perspektive kleiner und mittlerer Unternehmen eingegangen wird.
  • Thumbnail Image
    ItemOpen Access
    Distributed cooperative deep transfer learning for industrial image recognition
    (2020) Maschler, Benjamin; Kamm, Simon; Nasser, Jazdi; Weyrich, Michael
    In this paper, a novel light-weight incremental class learning algorithm for live image recognition is presented. It features a dual memory architecture and is capable of learning formerly unknown classes as well as conducting its learning across multiple instances at multiple locations without storing any images. In addition to tests on the ImageNet dataset, a prototype based upon a Raspberry Pi and a webcam is used for further evaluation: The proposed algorithm successfully allows for the performant execution of image classification tasks while learning new classes at several sites simultaneously, thereby enabling its application to various industry use cases, e.g. predictive maintenance or self-optimization.
  • Thumbnail Image
    ItemOpen Access
    Anwendungsfälle und Methoden der künstlichen Intelligenz in der anwendungsorientierten Forschung im Kontext von Industrie 4.0
    (2020) Maschler, Benjamin; White, Dustin; Weyrich, Michael
    Es wird erwartet, dass datengetriebene Methoden künstlicher Intelligenz im Kontext Industrie 4.0 die Zukunft industrieller Fertigung prägen werden. Obwohl das Thema in der Forschung sehr präsent ist, bleibt der Umfang der tatsächlichen Nutzung dieser Methoden unklar. Dieser Beitrag analysiert daher von 2013 bis 2018 veröffentlichte wissenschaftliche Artikel, um statistische Daten über den Einsatz von Methoden künstlicher Intelligenz in der Industrie zu gewinnen. Besonderes Augenmerk wird dabei auf die Trainings- und Evaluations-Datentypen, die Verbreitung in verschiedenen Industriezweigen, die betrachteten Anwendungsfälle sowie die geographische Herkunft dieser Artikel gelegt. Die resultierenden Erkenntnisse werden in praxisnahe Hinweise für Entscheider destilliert.
  • Thumbnail Image
    ItemOpen Access
    Deep learning based soft sensors for industrial machinery
    (2020) Maschler, Benjamin; Ganssloser, Sören; Hablizel, Andreas; Weyrich, Michael
    A multitude of high quality, high-resolution data is a cornerstone of the digital services associated with Industry 4.0. However, a great fraction of industrial machinery in use today features only a bare minimum of sensors and retrofitting new ones is expensive if possible at all. Instead, already existing sensors’ data streams could be utilized to virtually ‘measure’ new parameters. In this paper, a deep learning based virtual sensor for estimating a combustion parameter on a large gas engine using only the rotational speed as input is developed and evaluated. The evaluation focusses on the influence of data preprocessing compared to network type and structure regarding the estimation quality.
  • Thumbnail Image
    ItemOpen Access
    Realization of AI-enhanced industrial automation systems using intelligent Digital Twins
    (2020) Nasser, Jazdi; Ashtari Talkhestani, Behrang; Maschler, Benjamin; Weyrich, Michael
    A requirement of future industrial automation systems is the application of intelligence in the context of their optimization, adaptation and reconfiguration. This paper begins with an introduction of the definition of (artificial) intelligence to derive a framework for artificial intelligence enhanced industrial automation systems: An artificial intelligence component is connected with the industrial automation system’s control unit and other entities through a series of standardized interfaces for data and information exchange. This framework is then put into context of the intelligent Digital Twin architecture, highlight the latter as a possible implementation of such systems. Concluding, a prototypical implementation on the basis of a modular cyber-physical production system is described. The intelligent Digital Twin realized this way provides the four fundamental sub-processes of intelligence, namely observation, analysis, reasoning and action. A detailed description of all technologies used is given.
  • Thumbnail Image
    ItemOpen Access
    User-friendly, requirement-based assistance for production workforce using an asset administration shell design
    (2020) Al Assadi, Anwar; Fries, Christian; Fechter, Manuel; Maschler, Benjamin; Ewert, Daniel; Schnauffer, Hans-Georg; Zürn, Michael; Reichenbach, Matthias
    Future production methods like cyber physical production systems (CPPS), flexibly linked assembly structures and the matrix production are characterized by highly flexible and reconfigurable cyber physical work cells. This leads to frequent job changes and shifting work environments. The resulting complexity within production increases the risk of process failures and therefore requires longer job qualification times for workers, challenging the overall efficiency of production. During operation, cyber physical work cells generate data, which are specific to the individual process and worker. Based on the asset administration shell for Industry 4.0, this paper develops an administration shell for the production workforce, which contains personal data (e.g. qualification level, language skills, machine access, preferred display and interaction settings). Using worker and process specific data as well as personal data, allows supporting, training and instating workers according to their individual capabilities. This matching of machine requirements and worker skills serves to optimize the allocation of workers to workstations regarding the ergonomic workplace setup and the machine efficiency. This paper concludes with a user-friendly, intuitive design approach for a personalized machine user interface. The presented use-cases are developed and tested at the ARENA2036 (Active Research Environment for the Next Generation of Automobiles) research campus.