210 likes | 337 Views
Context Congregator: Gathering Contextual Information in CAPP. Umar Mahmud, Naima Iltaf Military College of Signals Farrukh Kamran Center for Advanced Studies in Engineering, Islamabad, Pakistan. Introduction.
E N D
Context Congregator: Gathering Contextual Information in CAPP Umar Mahmud, Naima Iltaf Military College of Signals Farrukh Kamran Center for Advanced Studies in Engineering, Islamabad, Pakistan
Context Congregator Introduction • Context-aware systems offer entirely new opportunities for application developers and for end users by gathering context data and adapting systems behavior accordingly. • High usability for mobile users. • Context Congregator Module gathers the low-level contextual data and then categorizes it as service context and user context, as part of the Context Aware Paradigm for Pervasive Computing Environment (CAPP), under research in MCS. • The gathered data is developed in Web Ontology Language (OWL) using the Protégé Tool. • This module is essentially a data structure for context information storage.
Context Congregator Related Work • The early context-aware system used only the location as the context • With time came the concept of a rich context that comprises of all the data that is involved in an interaction between a user and the service. • This data requires an easy to access as well as a well organized structure. • Different researchers have organized the data differently.
Context (Dey) Context (Schilit) Primary (Location, identification, time and activity) Secondary (All other information) User Computing Physical Context (Gwizdka) Context (Petrelli) Internal (State of user) External (State of environment) Material (Location, device and infrastructure) Social (Social aspects and personal traits) Context (Riaz) Context (Hofer) User Services System Physical (Measured through hardware) Logical (Specified by users or by monitoring user interactions) Context Congregator
Context Congregator Our Contribution • All of the context classifications are well defined but lack a structure that provides ease of access from the developer’s point of view. • More appropriately, in a pervasive environment interactions are typically between users and services. • Hence, we propose a structure that categorizes contextual data in terms of services’ and the users’ context. This has the effect of an easier selection of a service as requested by user on the basis of their contexts.
Context Congregator Context Models • Strang identifies five models to define and store context data in a machine process able form.
Context Congregator Context Models • The noticeable candidates for selection are the markup scheme model, object oriented model and the ontology based model. • The markup scheme model lacks the support of incomplete or ambiguous data while, the object oriented model does not provide reasoning to interpret the data. • The apparent choice is the ontology based model that is rich in representation, is an industry standard and provides inference techniques that maintains consistency of the structure.
Context Congregator Context Taxonomy • The contextual information must be classified into the users and the services categories. • The user interacts with the services through a number of devices. These devices are part of the user’s context since they are client to the services. • The parent concept is ‘Context’ that has sub concepts as ‘User Context’ and ‘Service Context’. • The user context is concerned with the user of the system including the human user as well as the devices held by the human user. • The service context is responsible for the context of the services. • The user and the service concepts are further sub classified into the temporal, activity, environmental, identification, spatial, community, health and advanced concepts.
Context Congregator Services Context • ServiceContext identifies only that contextual data which corresponds to the services in the system. • ServiceActivityInformation • This specifies the activity of the services in the system including the current status, length of task queue and the load on the service. • ServiceAdvancedInformation • Specifies the availability situation as when the service will be available for use due to heavy loads, the network bandwidth, bit rate and the battery power of the service (for mobile service). • ServiceCommunityInformation • The community of the services that includes the redundant, backup and similar services across multiple smart spaces as well as other services in the same smart space is identified in this concept. The status of the community is also kept in the sub concepts.
Context Congregator Services Context • ServiceEnvironmentInformation • The data identifying the environment in the vicinity of the service is part of the ServiceEnvironmentInformation concept. • ServiceIdentificationInformation • The identification parameters that include the name, ID, type, owner and the description are part of the ServiceIdentificationInformation concept. • ServiceSpatialInformation • Location of the service as well as the specification whether the service is mobile or fixed is recorded in ServiceSpatialInformation. • ServiceTemporalInformation • The time, day, date and season of the year information is specified here.
Context Congregator Users Context • The context of the user and the user devices are identified in the UserContext concept. • UserAdvancedInformation • The type of the devices held by the user and their status including the battery power and the activity, the emotional state of the user and the role are part of the UserAdvancedInformation concept. • UserAvtivityInformation • The current activity of the user as walking, talking, standing, sleeping, etc is part of this concept. This activity can be a composite value as a person maybe be talking while standing. • UserCommunityInformation • The community of the user includes the human peers in the vicinity, the devices of the human peers as well as other peer devices held by the same user. The activity status of the human users as well as the user devices in the surroundings is maintained in UserCommunityInformation.
Context Congregator Users Context • UserEnvironmentInformation • The environmental condition around the user is part of this concept.This includes It includes humidity, temperature, oxygen content, carbon dioxide content and dust level information. • UserHealthInformation • Physiological measurements of the user including the blood pressure, pulse, sugar level, body temperature, etc is part of the UserHealthInformation. This may include additional health constraints of the user like, high blood pressure, asthma etc. • UserIdentificationInformation • The identification parameters of the user including the address, name, and ID are specified by the UserIdentificationInformation.
Context Congregator Users Context • UserSpatialInformation • The location and mobility information of a user is identified here. If the user is mobile the expected location in addition to the current information can also be stored. • UserTemporalInformation • The time and date information of the user is specified in the UserTemporalInformation concept. It also includes the season of year information.
Context Congregator Representation Scheme • There are a number of languages that are used to implement ontology. • Ontology is the specifications of concepts or real world entities . • Ontology is like a description of the entities and there relationships that can exist for a community. • The ontology is expressed in a concrete formal notation called the ontology language. • There are a number of ontology languages both proprietary and standards- based. These include Cyc, KIF, DAML+OIL, RDF and OWL.
Context Congregator Web Ontology Language (OWL) • OWL is a standards-based markup language for publishing and sharing ontology on the Internet. • It is an extension to the Resource Description Framework (RDF) and is derived from the DARPA Agent Markup Language (DAML+OIL). • OWL is a part of the Semantic Web Vision where information on the web is machine readable, has exact meaning and is described. • OWL has an added feature that it allows strong reasoning support which makes it suitable for ontology based contextual models. • OWL provides a rich ontology representation and comes in three flavors OWL Lite, OWL DL and OWL Full. • We have selected OWL Full to represent the contextual data. OWL Full has the abilities of OWL Lite and OWL DL and is used to structure rich contextual data.
Context Congregator Controlled Vocabulary • The context congregator searches for sensor services present in the environment. • These services typically have non standardized service names. • A control vocabulary is a list of similar service names that is maintained in the context congregator. • Such a controlled vocabulary has keywords that provide us to search for a type of context that is to be sensed. • For example, a user asking for a ‘Temperature service’ could be replied by sensing the temperature through a ‘Weather service’. • This controlled vocabulary is presently static but it should be dynamic and adaptable to new instances. • Adaptation will be possible by determining similar services by identifying similar descriptions. • Services providing similar functions fall into the same category.
Context Congregator Context Features • Heterogeneity • OWL is based on XML standard and is an industry standard. • Vibrant context • The representation scheme provides hierarchy that supports quick access as compared to sequential access. This allows us to reflect vibrant and dynamic data changes • Privacy • Can be provided by encrypting externally • Distributed context • Data is gathered in a distributed environment. Missing data must be dealt with historical information. • Levels of interpretation • Hierarchy is provided • Consistency • Consistency has to be ensured externally
Context Congregator Conclusion • Contextual information is used to provide smart service delivery to a mobile and possibly naïve user thus providing context-awareness in pervasive environments. • The steps involved are the gathering of data, the representation of the gathered data and the subsequent interpretation of the data. • The context congregator provides a mechanism to gather low-level contextual data from the context sensors present in the system. • The context data is classified into users’ context and the services’ context. • The gathered data is then represented in OWL based ontology. • These two classifications of the context data provide a high-level of interpretation of the gathered low-level contextual data. • The interpreted data can be used to provide service discovery, service delivery or data depending on the user intention.
Context Congregator References • Dey, A.K. Understanding and Using Context. Personal and Ubiquitous Computing, 5, 1, 2001, 4-7. • Goslar, K. and Schill, A. Modeling Contextual Information using Active Data Structures. LNCS 3268, 2004. • Gruber, T. What is an Ontology? http://www.ksl.stanford.edu/kst/what-is-an-ontology.html • Hofer, T., Schwinger, W., Pichler, M., Leonhartsberger, G. and Altmann, J. Context-awareness on mobile devices – the hydrogen approach. In Proceedings of the 36th Annual Hawaii International Conference on Systems Sciences, (Hawaii, 2002). 292-302. • Horridge, M., Knublach, H., Rector, A., Stevens, R. and Wroe, C. A Practical Guide To Building OWL Ontologies Using The Protégé-OWL Plug-in and CO-ODE Tools Edition 1.0. http://protege.stanford.edu • Mahmud, U., Iltaf, N., Rehman, A., and Kamran F. Context-Aware Paradigm for a Pervasive Computing Environment (CAPP). WWW\Internet 2007, (Villa Real, Portugal, 2007). • Moran, T.P. and Dourish, P. Context-Aware Computing. Special Issue of Human-Computer Interaction, 16, 2001, 1-8. • Schilit, B. and Theiner. Disseminating Active Map Infrastructure to Mobile Host. IEEE Network, 8, 5, 1994, 22-32. • Strang, T. and Linnhoff-Popien, C. A Context Modeling Survey. First International Workshop on Advanced Context Modeling, Reasoning and Management, UbiComp 2004. • Petrelli, D., Not, E., Zancanaro, M., Strapparava, C. and Stock, O. Modeling and Adapting to Context. Personal and Ubiquitous Computing. 5(1). 2001, 20-24. • Weiser, M. The Computer for the 21st Century. Scientific American, 1991.