Semantic web technologies in pervasive computing: A survey and research roadmap

Juan Ye, Stamatia Dasiopoulou, Graeme Stevenson, Georgios Meditskos, Efstratios Kontopoulos, Ioannis Kompatsiaris, Simon Dobson
2015 Pervasive and Mobile Computing  
Pervasive and sensor-driven systems are by nature open and extensible, both in terms of input and tasks they are required to perform. Data streams coming from sensors are inherently noisy, imprecise and inaccurate, with di↵ering sampling rates and complex correlations with each other. These characteristics pose a significant challenge for traditional approaches to storing, representing, exchanging, manipulating and programming with sensor data. Semantic Web technologies provide a uniform
more » ... rk for capturing these properties. O↵ering powerful representation facilities and reasoning techniques, these technologies are rapidly gaining attention towards facing a range of issues such as data and knowledge modelling, querying, reasoning, service discovery, privacy and provenance. This article reviews the application of the Semantic Web to pervasive and sensor-driven systems with a focus on information modelling and reasoning along with streaming data and uncertainty handling. The strengths and weaknesses of current and projected approaches are analysed and a roadmap is derived for using the Semantic Web as a platform, on which open, standard-based, pervasive, adaptive and sensor-driven systems can be deployed. 1 http://www.w3.org/TR/owl-features/ 2 in OWL one can e↵ectively model and reason over taxonomic knowledge. This is a desirable feature in pervasive applications, where there is the need for modelling information at di↵erent levels of granularity and abstraction that will drive the derivation of further successively detailed contexts. Similarly, OWL supports consistency checking, another useful feature when dealing with imperfect context information coming from multiple sources. Resource Description Framework (RDF) The Resource Description Framework (RDF 2 ) provides a directed graph formalisation, with nodes representing resources and arcs representing properties. Its semantics are prescribed by two ontology languages: RDF Schema (RDFS ) and OWL. RDFS provides a basic vocabulary for dividing RDF resources into classes and introduces subClass and subProperty for capturing relations between classes and properties at varying levels of abstraction. On the other hand, OWL, as discussed later, provides a richer ontology language that supports expressing functional, transitive and inverse properties, equivalent properties and classes, and cardinality restrictions on the structure of class members. For sensor-driven systems, the benefits of these technologies emerge directly from their formality. RDF's use of Uniform Resource Identifiers (URIs) in identifying concepts and properties, combined with OWL's support for modelling equivalent classes and properties, allows determining whether lexically identical terms share the same meaning, or if two lexically di↵erent terms are synonyms or not. This formality has several advantages: (a) there is no single authority responsible for engineering ontologies or producing data; (b) entities may be described by combining concepts from di↵erent ontologies; (c) combining both ontologies and data from multiple sources is straightforward. Another benefit is domain-neutrality. RDF supports the representation of information across disparate application domains, unifying all data under a single model. Also, data across di↵erent components in a system and across di↵erent systems is seamlessly merged [26] . This can be contrasted to traditional database schemata, where terms and relations have no prescribed semantics, and XML Schema, which is concerned with the hierarchical structure of data elements and not with capturing the underlying relations. Technologies for managing RDF stores exist in the form of SPARQL 3 and SPARQL Update. SPARQL supports queries consisting of triple patterns, conjunctions, negations and disjunctions, while SPARQL Update supports the conditional insertion and removal of triples from an RDF store. Similarly to relational databases, there exist various tools supporting RDF-graph level manipulations or providing services to map RDF concepts to programming language type systems. Some representative examples are: Jena [84], OWL API [66] and RDFReactor [147]. A further RDF benefit is that an environment model can build upon, and interlink with, existing ontologybased domain knowledge through the principles of Linked Data -a set of best practices for exposing, sharing, and connecting pieces of knowledge on the SW [16] . The premise of Linked Data is that by using URIs and RDF to link to data sources, a semi-structured web of ontologically-represented data emerges that can be navigated and explored. Thus, Linked Open Data provides a structured means for accessing data from existing 'non-pervasive' sources and integrating it with existing RDF representations (or via an RDF wrapper for legacy data sources). This has particular potential for bootstrapping systems with the knowledge held by myriad social information sources on the web [117, 127]. OWL and OWL 2 OWL's design is strongly influenced by Description Logics (DL) [6] . DLs are a family of knowledge representation formalisms characterised by logically grounded semantics and well-defined reasoning. The main building blocks are concepts representing sets of objects (e.g. Person), roles representing relationships between objects (e.g. worksIn), and individuals representing specific objects (e.g. Alice). Starting from atomic concepts, such as Person, arbitrary complex concepts can be described through a rich set of constructors that define the conditions on concept membership. For example, the concept 9hasFriend.Person describes all those individuals that are friends with at least one person. 2 http://www.w3.org/RDF/ 3 http://www.w3.org/TR/sparql11-query/
doi:10.1016/j.pmcj.2014.12.009 fatcat:dehfs5inkbexpagk265co5tvxi