A framework for utilizing qualitative spatial relations between networked embedded systems

https://doi.org/10.1016/j.pmcj.2010.03.001Get rights and content

Abstract

With more and more everyday artifacts being equipped with networked embedded systems technology, their spatial situation or context–such as where they are located or whether two of them are near or far apart from each other–is becoming increasingly relevant. The emerging availability of sensor technologies for measuring properties of the physical space enables them to become aware of their spatial context and adapt to changes accordingly, which in turn contributes to the implementation of systems that operate autonomously in the background and interact with humans in a more unobtrusive way. In this article, we specifically address the use of spatial relations between technology-rich artifacts as well as their changes over time. A key aspect is the abstraction of spatial contexts in order to separate details which are not relevant for a certain application, and thereby save computational resources or provide spatial information in a way that is closer to human concepts of space. In this regard, our focus is on qualitatively represented spatial relations, which are used as the basic building blocks for the development of spatially aware applications. A novel software framework is presented for this purpose and evaluated with respect to its performance as well as its adequacy for building real-world applications.

Introduction

The miniaturization and seamless embedding of computing and wireless communication technologies into things of everyday use like tools, appliances and whole rooms lead to huge quantities of heterogeneous and interconnected smart objects, which are characterized by constrained user interfaces and limited computing resources. This development is mainly driven by technological advances within the past few decades, which have made it possible to shrink sensors and actuators as well as processing and wireless communication technologies to a size that enables their integration into virtually everything. A notable example for it is the recent integration of GPS and micro-electromechanical inertial sensors in a variety of consumer appliances like mobile phones, notebook computers, digital cameras and even wrist watches. Their density is expected to increase rapidly within the next decades, which will lead to a situation where humans are no longer able to manage and explicitly control them. Instead, it will be necessary to design and implement systems which operate autonomously, i.e. with as little human attendance as possible, and interact with humans in a more implicit and unobtrusive way. Clearly, the traditional approach of instructive systems [1] with their deterministic and context-free nature appears less appropriate as an architecture for providing services by the interaction of embedded and networked devices; instead, a more autonomous system architecture [2], [3] which implements context-dependent behavior is required.

For achieving autonomy, we consider two aspects to be especially important. The first one is the ability of an embedded system to acquire and interpret information about its environment from sensor data, which enables it to become aware of its context and adapt to changing contexts at runtime, and the second aspect is ad hoc context sharing, namely the spontaneous exchange of context information among systems within communication range. There has been much research on the former [4], [5], whereas location and spatial proximity have been the most often used information in context-aware applications. By sharing their contexts, it becomes possible for networked embedded systems to collaboratively adjust their actions to one another, and thus achieve coordinated behavior as has been demonstrated in several research projects [6], [7], [8]. According to [9], essential parts of the context of technology-rich artifacts are their spatial properties position, direction and extension as well as spatial relations between them; in the following, we refer to spatial properties and relations as spatial contexts. This is because they are part of the ever-changing physical world, and thus–due to movements or manipulations–also their spatial contexts are subject to frequent changes. The scope of this article is on the issue of making networked–and potentially mobile–embedded systems aware of their spatial context, with the aim to enable an adaptation to changing spatial situations and thus facilitate their contextual interaction in space. Our focus in this regard is on algebraic relations as well as their changes over time, which we consider especially relevant because of the fact that they describe the situation of a system with respect to its surroundings; therefore, they are fundamental for the development of applications whose functionality not only depends on the individual contexts of networked embedded systems, but also on the co-existence of other systems and their relations to each other. Such spatially aware applications are executed either on a single or on multiple embedded systems. Hence, we define spatial awareness as follows:

A system is spatially aware if it is able to autonomously determine and use its spatial properties, as well as to relate these properties to the spatial properties of other systems.

Spatial awareness requires means for acquiring spatial contexts from sensors, representing and reasoning (i.e. drawing conclusions) about perceived context information and sharing it among networked embedded systems in communication range. We use generic, structured “self-descriptions” for the latter, which encode the artifacts’ spatial contexts using an XML-based metadata format, and are exchanged among their embedded systems upon coming into communication range. A key aspect is the abstraction of context for its use by context-aware applications [10], [11]; for example, an application using location information may only be interested in high-level information like rooms and buildings instead of geographical coordinates. Such symbolic representations of locations have been addressed by several researchers (e.g. [12], [13], [14]), examples of abstracted positional relations–which are represented with meaningful names such as left or near instead of numeric angles or distances–can be found in [15]. The underlying idea is to provide context information in a sensor-independent way which (i) decouples low-level details that are not relevant for the application and (ii) allows programmers and developers to use context at a higher level of abstraction.

In order to represent spatial contexts, we propose two types of abstraction in the present work. First, quantitative abstractions are used for representing the spatial properties of technology-rich artifacts, which means that in a scene of artifacts just the spatial properties of those which are relevant for a spatially aware application are represented; numerical values are used to allow for calculating spatial relations between them. A concept referred to as Zones-of-Influence has been developed for this purpose, which builds upon initial work published in [16]. A Zone-of-Influence represents a geographic region, which is of relevance for and has an influence at a particular spatially aware application, by its numerical position, direction and spatial extension. Technology-rich artifacts are associated with one or more Zones-of-Influence at a time, together representing the spatial knowledge which is distributed across them in physical space and shared by an ad hoc exchange of self-descriptions. This allows artifacts to autonomously recognize spatial relations between their Zones-of-Influence, including distance and orientation relations as well as topological relations between their spatial extensions.

Second, qualitative abstractions are used for representing such spatial relations, which means that the continuous relation values are quantized according to the required accuracy and represented with discrete systems of symbols. Qualitative representations have attracted much interest by researchers in the field of spatial reasoning [17], [18]. Compared with quantitative approaches, they are often the preferred choice when the spatial cognition of humans is involved, which is due to the fact that (i) we are used to draw conclusions from coarse abstractions [19], [20] and (ii) qualitative approaches are similar to natural language terms and thus considered cognitively more adequate [21], [22], or when systems with limited computing resources or sensing capabilities are concerned [23], [24]. Hence, it should be easier for a programmer to develop applications by using high-level qualitative spatial relations whose semantics are adjusted to the application domain, and computationally less demanding reasoning techniques can be implemented due to the separation from details that are not relevant for a certain application or in a particular context.

In this article, we present novel models of spatial abstraction as well as an according software framework which is designed to run on embedded systems and facilitate the development of spatially aware applications. It is based on the concepts for recognizing and representing qualitative spatial relations between networked embedded systems mentioned above, and uses a rule-based approach for reasoning about spatial relations as well as their temporal relations to each other, which allows to trigger predefined actions upon observing certain patterns on the recognized relations. The architecture of the framework comprises layers for (i) the acquisition of spatial contexts from sensors and via an exchange of self-descriptions, (ii) the achievement of spatial awareness by recognizing, representing and reasoning about qualitative spatial relations, as well as (iii) the provision of services for the development of spatially aware applications. An essential aspect is that the relationship abstractions are encapsulated in components which can easily be exchanged at runtime, hence enabling the systems to adapt the availability of spatial relations and their semantics to the current application demands.

The paper is structured as follows. First, Section 2 presents fundamental aspects concerning the spatial awareness of wirelessly interconnected embedded systems as well as the exchange of context-information among them. Section 3 afterwards elaborates on the used spatial abstractions, namely (i) the quantitative representation of spatial properties with Zones-of-Influence and (ii) the qualitative representation of spatial relations. The framework is presented in Section 4, whereas an overview of the architecture is given and details about the implementation are presented. In this regard, the rule-based approach for reasoning about spatial relations and their changes over time is discussed in detail. Section 5 presents an evaluation of the framework performance, and Section 6 gives a comprehensive overview on related work. Section 7 finally concludes the paper, discusses the potentials of and lessons learnt from the presented concepts for realizing real-world applications, and gives an overview of open issues for future work.

Please note that this paper is intended to provide a comprehensive and concise presentation of our research on the use of spatial relations in spatially aware applications, which we have partly addressed in previous work. A fundamental aspect is the qualitative abstraction of space discussed in Section 3.2, which is rooted in initial work [25], [26] whose subject was the recognition and transitive inference of qualitative spatial distance and orientation relations between wireless embedded systems. The Zones-of-Influence concept (Section 3.1) and the ad hoc exchange of context information with structured self-descriptions (Section 2) build on previous work described in [16], [27], [28] respectively, which has been extended and adapted for the recognition of spatial relations between wireless embedded systems. First versions of the software framework and its rule-based reasoning approach have been presented in [29], [30], and they are described in thoroughly reworked and extended versions in Section 4.

Section snippets

Context awareness and spatial contexts

Context awareness refers to the ability of a system to acquire environmental information in order to understand its situation and adapt to changes accordingly [31], [32], [14]. It is considered important especially for applications where the environment changes frequently [5], [32], which is commonly the case in pervasive computing applications and requires an adaptation to the current context of use. From the different definitions of context awareness which appeared during the last 15 years,

Quantitative abstraction with Zones-of-Influence

Zones-of-Influence represent application-specific spatial regions which are associated with technology-rich everyday artifacts and shared among them by means of self-descriptions. This concept builds on work published in [16], where Zones-of-Influence are proposed as three-dimensional spatial regions that describe the space surrounding an artifact, and they serve as an explicit proximity model insofar as they are used for limiting the interaction to those systems whose Zones-of-Influence

A framework for spatial awareness

We have developed a flexible and modular software framework, with the aim to facilitate the implementation of spatially aware applications for mobile and ad hoc networked embedded systems. Instead of having to re-implement the recognition and maintenance of spatial relations for each application in which they are required, the framework allows to reuse existing implementations and dynamically adapt the calculation and use of relations to the current application demands. It runs on each embedded

Performance evaluation

The presented framework allows to build spatially aware applications using different numbers of Zones-of-Influence, spatial relations and rules. In this section, the results of performance tests are presented, which have been carried out to show the runtime efficiency and scalability of the framework. Scalability is a key issue for building complex applications which comprise a multitude of interacting embedded systems, as the required processing time of each system increases with (i) the

Related work

This section finally gives an overview on related work addressing the use of qualitative spatial relations by networked embedded systems, which is–in contrast to location sensing–a sparsely investigated topic in the field of pervasive computing. A closely related project is the project Relate [73], [36], [15], which uses specialized USB dongles for providing mobile devices with an awareness about both quantitative and qualitative spatial relations to other co-located devices. Compared to our

Conclusion

In future computing systems, which are expected to be increasingly interconnected and embedded in the environment, operate more autonomously and with which humans will interact in a more natural and implicit way, the consideration of spatial contexts–and in particular of spatial relations between co-located systems–will play a significant role. It enables embedded systems to adapt to their spatial situation, for which purpose they must be able to sense their context, process and reason about

Acknowledgements

This work is supported by the FP7 ICT Future Enabling Technologies programme of the European Commission under grant agreement No. 225938 (OPPORTUNITY). Parts of it have been developed in a research project with Siemens AG Germany, Corporate Technology, CT SE 2.

Clemens Holzmann received master’s degrees in both Computer Science (2004) and Business Informatics (2006) from the Johannes Kepler University Linz, where he also received the doctorate in Computer Science (2008). From 2004 to 2010, he was with the Department of Pervasive Computing at the Johannes Kepler University Linz at the level of an assistant professor. During this time, he was involved in the EU project OPPORTUNITY and in several research projects funded by SIEMENS Germany. In 2010, he

References (85)

  • N. Kohtake et al.

    u-Texture: Self-organizable universal panels for creating smart surroundings

  • A. Dey

    Understanding and using context

    Personal and Ubiquitous Computing

    (2001)
  • D. Salber et al.

    The context toolkit: Aiding the development of context-enabled applications

  • A. Ferscha et al.

    Context-awareness for group interaction support

  • U. Leonhardt, Supporting location-awareness in open distributed systems. Ph.D. thesis, Imperial College London, May...
  • W.N. Schilit, A system architecture for context-aware computing, Ph.D. thesis, Columbia University, May...
  • G. Kortuem et al.

    Sensing and visualizing spatial relations of mobile devices

  • A.G. Cohn et al.

    Qualitative spatial representation and reasoning: an overview

    Fundamenta Informaticae

    (2001)
  • J.R. Hobbs et al.

    Spatial representation and reasoning

  • K.D. Forbus

    Qualitative reasoning

  • A. Musto, Qualitative Repräsentation von Bewegungsverläufen. Ph.D. thesis, Technical University Munich, August...
  • J. Renz et al.

    Qualitative direction calculi with arbitrary granularity

  • T. Tenbrink, Semantics and application of spatial dimensional terms in English and German. Tech. Rep. SFB/TR 8 Report...
  • C. Freksa

    Using orientation information for qualitative spatial reasoning

  • B. Gottfried et al.

    Spatiotemporal reasoning for smart homes

  • C. Holzmann

    Inferring and distributing spatial context

  • C. Holzmann et al.

    Towards collective spatial awareness using binary relations

  • A. Ferscha et al.

    Building flexible manufacturing systems based on Peer-its

    EURASIP Journal on Embedded Systems

    (2007)
  • A. Ferscha et al.

    Digital aura

    Proceedings of the 4th Int’l Conference on Pervasive Computing, Pervasive 2004

    Advances in Pervasive Computing

    (2004)
  • C. Holzmann

    Rule-based reasoning about qualitative spatiotemporal relations

  • C. Holzmann et al.

    Using spatial abstractions in industrial environments

  • J. Pascoe

    The stick-e note architecture: Extending the interface beyond the user

  • B. Schilit et al.

    Context-aware computing applications

  • P.J. Brown et al.

    Context-aware applications: from the laboratory to the marketplace

    IEEE Personal Communications

    (1997)
  • R. Want et al.

    The active badge location system

    ACM Transactions on Information Systems

    (1992)
  • M. Hazas et al.

    A relative positioning system for co-located mobile devices

  • A.I. Abdelmoty et al.

    An intersection-based formalism for representing orientation relations in a geographic database

  • F. Hupfeld et al.

    Spatially aware local communication in the RAUM system

  • M. Strohbach et al.

    Cooperative artefacts: Assessing real world situations with embedded technology

  • H. Lei et al.

    The design and applications of a context service

    Mobile Computing and Communications Review

    (2002)
  • M. Román et al.

    A middleware infrastructure for active spaces

    IEEE Pervasive Computing

    (2002)
  • C. Portele, OpenGIS Geography Markup Language (GML) encoding standard, version 3.2.1. Open Geospatial Consortium Inc....
  • Cited by (10)

    • FIA<inf>5</inf>: A customized Fuzzy Interval Algebra for modeling spatial relevancy in urban context-aware systems

      2014, Engineering Applications of Artificial Intelligence
      Citation Excerpt :

      Context appears as a fundamental key to enable systems to filter relevant information from what is available (Dey, 2001; Schilit et al., 1994; Siewe et al., 2009; Cook and Das, 2012), to choose relevant actions from a list of possibilities (Coronato and Pietro, 2011; Hong et al., 2009; Chedrawy and Abidi, 2006), or to determine the optimal method of information delivery (Pan et al., 2007). The major challenge of context-aware systems is the separation of the relevant from the irrelevant information (Holzmann and Ferscha, 2010; Choi et al., 2012; Raper et al., 2002; Grossmann et al., 2005; Reichenbacher, 2007; Thyagaraju and Kulkarni, 2012). Among all types of relevancies, spatial relevancy is the main type which could handle other types of relevancies such as user preferences, history, etc. (Afyouni et al., 2011).

    • Grounding ecologies on multiple spaces

      2012, Pervasive and Mobile Computing
      Citation Excerpt :

      In addition, our concrete framework can be used as a generic service to support context awareness [33] as organisms could utilize its features to acquire and disseminate context information. Finally, models and frameworks have been recently proposed with the aim of obtaining spatially localized information (see for instance [34,35]). In contrast with our approach, they emphasize physical spaces only and they do not propose space-aware interaction mechanisms.

    • A framework for qualitative representation and reasoning about spatiotemporal patterns

      2014, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    View all citing articles on Scopus

    Clemens Holzmann received master’s degrees in both Computer Science (2004) and Business Informatics (2006) from the Johannes Kepler University Linz, where he also received the doctorate in Computer Science (2008). From 2004 to 2010, he was with the Department of Pervasive Computing at the Johannes Kepler University Linz at the level of an assistant professor. During this time, he was involved in the EU project OPPORTUNITY and in several research projects funded by SIEMENS Germany. In 2010, he joined the Upper Austria University of Applied Sciences, where he is now a professor in the degree programme Mobile Computing. His research interests include Pervasive and Mobile Computing, Embedded Systems and Human–Computer Interaction.

    Alois Ferscha was with the Department of Applied Computer Science at the University of Vienna at the levels of assistant and associate professor (1986–1999). In 2000 he joined the University of Linz as full professor where he heads the Excellence Initiative ”Pervasive Computing”, the department of Pervasive Computing, and the Research Studio Pervasive Computing Applications. He is active in international EU funded projects (FET FP7: PANORAMA, SOCIONICAL, OPPORTUNITY, BeyondTheHorizon, InterLink, CRUISE) as well as national (funded projects SPECTACLES, PowerSaver, WirelessCampus, MobiLearn) research, and keeps tight cooperations with industrial stakeholders (SIEMENS Project FACT, IBM Project VRIO). Currently he is focused on Pervasive and Ubiquitous Computing, Embedded Software Systems, Wireless Communication, Multiuser Cooperation, Distributed Interaction and Distributed Interactive Simulation.

    View full text