Towards a novel learning assistant for networked automation systems

Abstract

Due to increasing requirements on functionality (e.g. self-diagnosis, self-optimization) or flexibility (e.g. self-configuration), future automation sys-tems are demanded to be more and more intelligent. Therefore the systems are desired to learn new knowledge from other systems or its environment. The pur-pose of this work is to propose a prospective concept of learning assistant for networked automation systems. With the help of the assistant, an automation sys-tem can obtain new knowledge by collaborating with other systems to improve its prior knowledge. So that the system user is liberated from continuously providing new knowledge to an automation system. 1 Introduction and Motivation Nowadays, industrial automation systems are always demanded more functionality (e.g. self-diagnosis, self-optimization) or flexibility (e.g. self-configuration). In this case, they are developed to be more and more complex with increasing integration of information and communication technology. The situation becomes even more com-plex, when the automation systems are networked together. Therefore, industrial auto-mation systems are more difficult to be operated than before. In order to ease the oper-ability of these systems, assistance systems are introduced to support system users, who can then neglect the complexity of the systems. The intelligence of assistant systems lies in their knowledge base, which has been predefined by the system developer. The knowledge is nothing more than a collection of facts, events, beliefs, and rules, organized for systematic use [1]. There has been a lot of research on intelligent assistants. Most of the concepts have the following two features: (1) the intelligent assistants are always developed for an individual automation system which is not connected to other systems; (2) the prior knowledge of the system will not be changed or only changed in a small scope. With the emerging of industry 4.0, automation systems are required to connect with each other and coordinate themselves in different responsibilities. In this context, the intelligent assistant should be improved with the consideration of a networking envi-ronment and learning new knowledge from other systems in the same network. This art O. Niggemann, J. Beyerer (eds.), Machine Learning for Cyber Physical Systems, Technologien für die intelligente Automation 1, DOI 10.1007/978-3-662-48838-6_7, © Springer-Verlag Berlin Heidelberg 2016