In this paper we introduce online semi-supervised growing neural gas (OSSGNG), a novel online semi-supervised classification approach based on growing neural gas (GNG). Existing semi-supervised classification approaches based on GNG require that the training data is explicitly stored as the labeling is performed a posteriori after the training phase. As main contribution, we present an approach that relies on online labeling and prediction functions to process labeled and unlabeled data uniformly and in an online fashion, without the need to store any of the training examples explicitly. We show that using on-the-fly labeling strategies does not significantly deteriorate the performance of classifiers based on GNG, while circumventing the need to explicitly store training examples. Armed with this result, we then present a semi-supervised extension of GNG (OSSGNG) that relies on the above mentioned online labeling functions to label unlabeled examples and incorporate them into the model on-the-fly. As an important result, we show that OSSGNG performs as good as previous semi-supervised extensions of GNG which rely on offline labeling strategies. We also show that OSSGNG compares favorably to other state-of-the-art semi-supervised learning approaches on standard benchmarking datasets.

The manufacturing of individualized products down to batch size 1 poses ongoing challenges for the design and integration of future production systems. Today's production lines with a high degree of automation achieve high efficiency, but usually come with high costs for adaptation to product variants. In order to combine full automation with high flexibility, we propose a concept for the dynamic composition of automation components in a modular production system that facilitates the rapid adaptation of collaborative and robotsupported manufacturing processes. To achieve this, we integrate self-descriptive automation components at runtime into the control architecture of the production system using a Plugand-Produce approach. While the location and orientation of automation components in the modular production system are derived from physical human-robot interaction, the adaptation and verification of the robot behavior is made possible through a simulation-based planning subsystem. Once this dynamic reconfiguration process by the machine setter is finished, the adapted production process is executed in a fully automated way with high efficiency. A case study carried out in an industrial collaboration project on flexible assembly demonstrates the benefits of the presented approach.

No abstract

Abstract. Growing neural gas (GNG) has been successfully applied to unsupervised learning problems. However, GNG-inspired approaches can also be applied to classification problems, provided they are extended with an appropriate labelling function. Most approaches along these lines have so far relied on strategies which label neurons a posteriori, after the training has been completed. As a consequence, such approaches require the training data to be stored until the labelling phase, which runs directly counter to the online nature of GNG. Thus, in order to restore the online property of classification approaches based on GNG, we present an approach in which the labelling is performed online. This online labelling strategy better matches the online nature of GNG where only neurons -but no explicit training examples -are stored. As the main contribution, we show that online labelling strategies do not deteriorate the performance compared to offline labelling strategies.

Der Beitrag beschreibt ein Hard- und Softwarekonzept für vernetzte Fertigungsmodule. Eine modulare Systemarchitektur sowie die dezentrale Steuerung durch Prozessmodelle auf Basis von BPMN2 erlauben eine kundenspezifische Produktion bis hin zu Losgröße eins. Anhand eines vertikal in die Unternehmens-IT integrierten Demonstrators wurden die Vorteile als Showcase für Industrie 4.0 auf verschiedenen Fachmessen erlebbar. Der innovative Ansatz wurde im Verbundprojekt itsowl-FlexiMon im Rahmen des BMBF Spitzenclusters „Intelligente Technische Systeme OstWestfalenLippe“ (it’s OWL) entwickelt.   This contribution describes a distributed modular production system for individualized production. A modular system architecture and semi-autonomous cell control based on executable process models with BPMN2 are used to realize a customer specific production down to lot size one. The advantages have become tangible through a vertically integrated demonstrator that has been exhibited at different fares and showcases the progress towards Industry 4.0. The overall approach was developed in the project itsowl-FlexiMon in the framework of the BMBF leading edge cluster „Intelligent Technical Systems OWL“ (it’s OWL).