Service Oriented Computing to Self-Learning Production System M.K. Uddin, A. Dvoryanchikova, J.L. Martinez Lastra Tampere University of Technology, Finland [mohammad.uddin, aleksandra.dvoryanchikova, jose.lastra]@tut.fi S. Scholze, D. Stokic Institut für angewandte Systemtechnik Bremen, Germany [scholze, dragan]@atb-bremen.de G. Cândido, J. Barata UNINOVA - Instituto de Desenvolvimento de Novas Tecnologias, Potugal [gmc, jab]@uninova.pt Abstract- The aim of this manuscript is to present what is Self- Learning production system and how service oriented architecture (SOA) and supporting technologies are bridged together to implement this new concept in the ongoing EU Self- Learning production system project. A brief review of the most recent EU projects that have reported results relevant to the main discussed investigation problems is presented. Reference architecture and functionalities of Self-Learning production system is introduced aiming for improved control and maintenance in production plants. Service oriented computing to Self-Learning production system is proposed to meet the required level of flexibility, interoperability and communications needs for reusable Self-Learning services. A roadmap for future research is defined. I. INTRODUCTION Technologies leveraging artificial intelligence at the factory floor, knowledge based system development and machine learning are being studied to make capabilities of self-X properties like self-adaptation, self-optimization, self- evolution and self-maintenance in production systems. This manuscript addresses Self-Learning production system, which is a new concept to apply cybernetic principles to derive intelligent production systems. The system self- adapt and learn in response to the dynamic changes in contextual information extracted from all factory levels. Context awareness approach addresses the integration of control and maintenance processes for necessary adaptation, which results in maintenance cost reduction and improves the overall equipment effectiveness especially regarding system availability and productivity. A reliable and secure software service based integration infrastructure using distributed networked embedded services in device space is the key to achieve such system. The main purpose of this work is to present what is Self- Learning production system and how SOA as an architectural paradigm and supporting technologies are bridged together to achieve a seamless enterprise wide connectivity using flexible, loosely coupled and reusable Self-Learning services. The contribution of this manuscript is three fold. Firstly, a brief review of the related research and exploitable results reported in the most recent EU projects are presented (Section II); and the bridging of SOA to the manufacturing world is addressed (Section III). Secondly, the concept of Self-Learning production system is presented introducing the generic reference architecture. The functionalities of the architectural components are also described (Section IV). Finally, this work addresses service oriented computing to Self-Learning production system. The focus is to define a SOA-based communication infrastructure to provide universally accepted set of interoperability standards for building, describing, cataloguing and managing reusable Self- Learning services (Section V). The future research roadmap is outlined in section VI and the conclusions are drawn in section VII. II. RELATED WORKS To enable self-‘X’ features in production systems, researchers have considered modeling of cognitive behavior, control behavior and reactive behavior. Control strategy modeling using CAE tools, reactive/proactive behavior modeling with real time variants of state machines and cognitive behavioral modeling with UML based approaches is addressed in [1] to realize self-optimizing mechatronic systems. CAE tool CAMeL [2] and CASE tool Fujaba [3] is integrated at tool level and UML extension at the semantic level is addressed to achieve semantically rich interfaces at the component level to deal with multi agent systems. CRC 614 (Collaborative Research Center) [4] research addresses self-optimizing design and implementation issues on future electro-mechanical products having inherently partial intelligence. The primary objectives are the scientific exploration of self-optimizing principles on engineering products, design methods and required tools development. The aim is the realization of the principles on the level of hardware, system software and control software, which goes beyond the traditional rule based and adaptive control strategies. The results are further extended to design of systems, which are flexible, react autonomously and take actions in changing operating conditions. Principles of autonomous and adaptive control in assembly systems are addressed in [5] focusing on intelligent computational methods, agent based systems and reconfigurable manufacturing strategies. The goal is to utilize these principles to bring flexibility down to the factory floor. Architectural challenges of a self-managed system with a