An autonomic ontology-based approach to manage information in home-based scenarios: From theory to practice

Highlights

• Ontologies and the autonomic computing paradigm are proposed for integration in telemonitoring scenarios.

• The expressiveness of SPARQL is put into practice.

• Technical management was successfully studied with this approach.

Abstract

Data integration is a key challenge in scenarios where different sources of information form part of a common service, as is the case in home-based scenarios. Moreover, an integrated management procedure is deemed necessary in order to unify the ways of accessing, controlling, evaluating and transferring information remotely. In this paper, a combination of the autonomic computing paradigm with ontologies is proposed to meet this challenge. The potential solution has been studied from both theoretical and practical points of view. First, an ontology model driven approach based on the autonomic computing paradigm (MAPE: monitor, analyze, plan and execute) has been proposed as a reference knowledge framework to unify both managed data and management procedures. Secondly, the proposed knowledge model has been applied to the management of technical information in home-based telemonitoring scenarios. An agent to perform the management tasks described in the ontology has been implemented and has been successfully evaluated in terms of its application to the definition of different actions within a telemonitoring scenario. The solution presented in this work offers a generic solution for the integration of data and its management procedure in home-based scenarios.

Introduction

Data integration has emerged in recent decades as an interesting research challenge. This interest has been mainly motivated by the need for seamless access to data provided by heterogeneous information sources that are expected to work together (exchanging and integrating information) to provide a common service [1], [2]. This interoperability problem occurs in home environment scenarios where a variety of heterogeneous devices having their own formats to represent and exchange data are used together to provide users with services such as patients' telemonitoring or home-automation applications [3], [4], [5]. Some of the devices within a typical home scenario such as sensors and/or medical devices (MDs) transfer the acquired data to a concentrator device (designated in Fig. 1 as HG, Home Gateway) whose main purpose is to collect information from the different data sources. In general, in a typical home scenario all these devices are part of an architecture that usually includes two sites: the home site and a remote site (where an external manager (EM) is located) linked by a communication network (see Fig. 1). At the home site, data provided by the different devices and also sometimes user feedback are transferred and collected in the HG. Subsequently, the data collected in the HG is generally transferred to the remote site or needs to be processed in order to extract additional information. Hence, apart from data integration, management tasks also need to be integrated in the HG. In the remote site, a server device designated as the Telemonitoring Server (TS) is commonly used to manage information provided and collected from all the HGs linked to it. In summary, in addition to providing a common understanding of the terminologies and concepts used, it is as important to provide the HG with common knowledge about how actions should be performed to collect, process and remotely access data. Although in recent years standards have been proposed to address these integration and interoperability challenges at different points of this communication scenario (e.g. the ISO/IEEE 11073, X73, for MD communication could be used at the home site) [6], a higher level of abstraction is required in order to provide more than a common format to exchange data. A common knowledge to deal with both integrations is needed, thus relating all the sources that take part in the management process. Hence, the development of knowledge oriented frameworks is thus necessary to address this challenge and consequently provide effective and shareable services.

Such integration in the HG should be addressed from both theoretical and practical points of view [7]. First, from a theoretical point of view, a formal model is required in the HG to achieve a common understandable knowledge for all the sources that share and exchange data not only at the home site but also at the remote site. Hence, ontologies constitute the perfect theoretical approach to deal with data integration in the proposed scenario. They provide a knowledge oriented framework in which different data sources can be easily mapped and, in contrast to other data structure technologies, they provide mechanisms to reason upon data defined in the model. In fact, three popular environments where ontology-based approaches have been successfully applied to solve interoperability problems are clinical environments [7], [8], [9], [10], [11], [12], home-smart environments [13], [14], [15] and network management applications [16], [17]. The development of a specific tool to manipulate and map incoming data into the new model is required. Apart from developing the theoretical model, practical applications to integrate data are required [18], [19].

It should be noted that the HG element works in a dynamic environment where over time new devices can be added to or removed from it. In order to avoid continuous human intervention to manage applications supported by the HG, clearly some kind of self-management capabilities should be introduced into the HG (e.g. to perform updating tasks and provide services according to the configured scenario). Hence, this scenario seems to be an appropriate area for the application of some ideas based on the so-called autonomic computing paradigm [20]. According to this paradigm, an autonomic element implements an intelligent control loop that consists of collecting data, analyzing the data, determining a plan or sequence of actions in response to the analysis and finally executing the plan/sequence of actions. Consequently, the combination of an ontology-based approach and the autonomic computing paradigm is proposed in this paper to deal with the data and management procedure integration challenge in home-based scenarios.

The idea proposed and implemented is to develop an ontology to clearly describe the managed data and explicitly describe in the data the representation of the management procedures, i.e. the tasks that comprise the control loop explained above. Furthermore, rules are proposed to activate and relate tasks described in the ontology. Hence, by defining instances of the proposed ontology and rules, managed elements are provided with a profile where all the tasks to be performed for the management procedure are clearly described and personalized management services can be provided. It is argued by the authors that the proposed approach constitutes a formal and generic solution that can be used not only to represent but also to manage any type of information in a home-based scenario when data provided by the different sources are collected and managed in the HG. In order to show the practical applicability of this model, a specific case study using the proposed approach for technical management in home-based telemonitoring scenarios is presented in detail. To support this application, an agent has been developed to host the ontology and execute the tasks in the HG. The paper thus addresses both the theoretical aspects and the practical applicability of the proposed approach.

The remainder of the paper is structured as follows: Section 2 provides background information about the technologies involved in the solution and related work is also discussed. Section 3 describes in detail the theoretical ontology model proposed. Section 4 addresses the case study and presents the final ontology model and the prototype performance. Finally, the conclusions are set out in Section 5.