Fraunhofer-Institut für Optronik, Systemtechnik und Bildauswertung IOSB

Filters

3
2 1
Reset filters

Settings
Now showing 1 - 3 of 3
  • Publication
    Towards Real-World Deployment of Reinforcement Learning for Traffic Signal Control
    ( 2021)
    Müller, Arthur
    ;
    Rangras, Vishal
    ;
    Ferfers, Tobias
    ;
    Hufen, Florian
    ;
    Schreckenberg, Lukas
    ;
    Jasperneite, Jürgen
    ;
    Schnittker, Georg
    ;
    Waldmann, Michael
    ;
    Friesen, Maxim
    ;
    Wiering, Marco
    Sub-optimal control policies in intersection traffic signal controllers (TSC) contribute to congestion and lead to negative effects on human health and the environment. Reinforcement learning (RL) for traffic signal control is a promising approach to design better control policies and has attracted considerable research interest in recent years. However, most work done in this area used simplified simulation environments of traffic scenarios to train RL-based TSC. To deploy RL in real-world traffic systems, the gap between simplified simulation environments and real-world applications has to be closed. Therefore, we propose LemgoRL, a benchmark tool to train RL agents as TSC in a realistic simulation environment of Lemgo, a medium-sized town in Germany. In addition to the realistic simulation model, LemgoRL encompasses a traffic signal logic unit that ensures compliance with all regulatory and safety requirements. LemgoRL offers the same interface as the well-known OpenAI gym toolkit to enable easy deployment in existing research work. To demonstrate the functionality and applicability of LemgoRL, we train a state-of-the-art Deep RL algorithm on a CPU cluster utilizing a framework for distributed and parallel RL and compare its performance with other methods. Our benchmark tool drives the development of RL algorithms towards real-world applications.
  • Publication
    Investigation in IoT and 5G architectures for deployment of Artificial Intelligence into urban mobility and production
    ( 2021)
    Ferfes, Tobias
    ;
    Müller, Arthur
    ;
    Schriegel, Sebastian
    ;
    Jasperneite, Jürgen
    The automation industry discusses the deployment of processing resources in applications since decades. The two major subjects of discussion are centralized and decentralized computing. The deployment of artificial intelligence (AI) applications and the processing power is facing the same issue nowadays. AI is used in many different areas like smart city, industrial automation for pattern recognition or as an expert system and so on. However, it remains a central question how to deploy the AI within an application and how to connect the AI to other devices, services and agents. In this paper we discuss state-of-the-art architecture to structure general IoT applications, as well as, the two most common processing concepts. Furthermore, use cases in urban mobility and industrial automation to deploy AI applications are presented and an In-depth view to the implementation is discussed. Pros, cons and problems of the implementations are shown, which can provide ideas and references when deploying an AI application.
  • Publication
    An OPC UA based approach for dynamic-configuration of security credentials and integrating a vendor independent digital product memory
    ( 2014)
    Blume, Marco
    ;
    Koch, Nils
    ;
    Imtiaz, Jahanzaib
    ;
    Flatt, Holger
    ;
    Jasperneite, Jürgen
    ;
    Schleipen, Miriam
    ;
    Sauer, Olaf
    ;
    Dosch, Steffen
    This paper presents an approach to securely integrate industrial devices into automation systems with a minimal engineering effort. A special specific focus is on the needed communication architecture that is based on the platform independent and vendor neutral technology OPC UA. The paper also describes the need of a digital product memory besides a life cycle data harvesting to facilitate such seamless integration; this is by means of presenting semantics of operations to an external system. As part of the work, a case study has been identified; different architectural aspects are evaluated and essential system components are realized/implemented/integrated as a proof of concept. Principle results include the implementation of a BeagleBone Black-based Secure Plug & Work I/O field device with an extended real-time industrial communication interface and a semantically enriched OPC UA server that provides vendor neutral configuration and an I/O data service interface. Furthermore, the result provides a platform independent and standardized way to represent a field device to external systems, to enable intelligent technical systems to communicate and orchestrate a seamless and secure integration.