Context-aware knowledge-based middleware for selective information delivery in data-intensive monitoring systems

Abstract

Multiple embedded devices in modern control and monitoring systems are able to sense different aspects of the current context such as environmental conditions, current processes in the system and user state. The number of captured situations in the environment and quantity and variety of devices in the system produce considerable amounts of data, which should be processed, understood and followed by corresponding actions. However, fully delivered to the user regardless of their role in the system and needs, data flows cause cognitive overload and thus may compromise the safety of the system depending on the timely response of the operators. This paper addresses the problem of selective information delivery with respect to the user׳s role in the system, his needs and responsibilities, by proposing context-aware information management middleware. The system utilizes Semantic Web technologies by capturing relevant information in the knowledge model of the system, which decouples data from the application logics. A clear division of data and application logics enables context-awareness and facilitates the reconfiguration process, when new information should be added into the system. The chosen approach is justified with an analysis of main trends in context-aware solutions. The engineering principles of the knowledge model are described and illustrated with simple scenarios from the building automation domain. The prototype developed proves the feasibility of the approach via performance evaluation and demonstrates the reconfiguration capabilities of information flows in the system. Further work assumes the extension of the knowledge model and integration of the system with adaptive human–machine interfaces for multi-role and multi-user environments.

Introduction

The evolution of control and monitoring systems has enabled information accessibility for almost any aspect of the environment (Sauter et al., 2011). In modern systems it is possible to integrate large numbers of various sensors to capture environmental conditions (e.g. temperature, humidity, and light), particular features of the process of interest (e.g. the product status in the production line) and the state of the users involved in the process or system operation. Potentially, it could offer advantages for timely situational awareness of the all system users. In reality growing information complexity challenges the usability of the system, human performance, and safety overwhelming the cognitive capacities of a user. This paper addresses a problem of selective information delivery to different users of the system by proposing context-aware information management middleware (IMM), which is able to analyze the run-time data of the system and deliver the information which is relevant to the user given the current situation and user needs/responsibilities.

The aim is to bring the right information to the right user at the right time, thus information management middleware should be able to determine the “rightness” according to the situation. The introduction of context-awareness is claimed to enable the desired flexibility of system behavior (Baumgartner et al., 2010, Fischer, 2012). The context can be defined as “any information that can be used to characterize the situation of an entity, where an entity can be a person, place, or physical or computational object” (Abowd and Dey, 1999), and then the system is context-aware “if it uses context to provide relevant information and/or services to the user, where relevancy depends on the user׳s task” (Dey, 2001). Both definitions are quite generic and allow different technological solutions for context-aware systems providing the right services (Emmanouilidis et al., 2013), actions (Feng et al., 2009) or system adaptation.

This paper proposes context-aware information management middleware for data intensive monitoring systems, where user is exposed to numerous sources of data coming from heterogeneous devices of the system and various situations of the environment. The middleware provides the user with the relevant information depending on the current situation, user role in the system, his needs and responsibilities. The main technological trends of achieving context-awareness in the systems are presented and analyzed, and the chosen approach using Semantic Web technologies (specifically OWL ontologies, SPARQL and SPARQL/Update¹ queries) is justified. The system utilizes a reconfigurable knowledge model (ontologies) which decouples data from the application logic. The clear division of data and application logic enables context-awareness and (1) facilitates the system reconfiguration process when new information is introduced into the system; (2) enables the implementation of domain independent tools which facilitate the process of adapting IMM in new domain. The proposed approach is supported by ontological engineering principles, a description of a prototype already implemented, and describing reference cases from the building automation domain.

The rest of the paper is organized as follows. Section 2 overviews the state of the art for context-aware systems and justifies the choice of Semantic Web technologies for the solution proposed. Section 3 describes the functional architecture of the system and the enabling technologies for the main architectural blocks. Section 4 states the main design principles of the knowledge model of the system, the query templates enabling reconfiguration and the domain independent capabilities of the solution. Section 5 describes the tools for implementing IMM in detail and is followed by section 6 with simple examples of scenarios from the building automation domain exemplifying the knowledge model and demonstrating in function the main principles of the proposed solution. Section 7 presents performance evaluation of developed prototype; and finally, Section 8 concludes the paper and highlights directions for future work.