Ubiquitous Semantic Space: A context-aware and coordination middleware for Ubiquitous Computing

Abstract

Ubiquitous computing poses the challenge of increased communication, context-awareness and functionality. In a highly dynamic and weekly connected ubiquitous environment, continuous access to the network (synchronous communication) is very difficult. So it's necessary to go for tuple space which provides asynchronous communication without any loss in data. Tuple space offers a coordination infrastructure for communication between autonomous entities by providing a logically shared memory along with data persistence, transactional security as well as temporal and spatial decoupling properties that make it desirable for distributed systems [2] such as ubiquitous computing. In order to automate the task and the system to be intelligent, context awareness is a must. This can be achieved by using semantic web technology. Existing middleware's for ubiquitous computing concentrates on RPC communication paradigm and deals with context-awareness separately. In our approach of constructing the middleware we provide common solution to both communication and context-awareness using ubiquitous semantic space. Ubiquitous semantic space [5] brings together tuple space, semantic web technologies and ubiquitous computing. Hence in this paper, we introduce a context-aware and co-ordination middleware framework for ubiquitous environment using ubiquitous semantic space. Ubiquitous semantic space uses ontologies to define the semantics of various concepts. Using ontologies facilitates different agents in the environments to have a common semantic understanding of different contexts. Ontology is represented using ontology web language, OWL [6]. We have modeled a ubiquitous semantic space ontology structure suitable for communicating conceptual information among the agents. Our model also incorporates context-triggered action which is more useful for real-time ubiquitous application having reactive behavior. For enabling context-triggered action, our model has reactive space into ubiquitous semantic space. Reactive space, stores rules written in SWRL [22], semantic web rule language and fired using JESS [3] reasoner at the appropriate time. The middleware could easily adapt to changes in the environment. The structure of the ubiquitous semantic space is designed in a fashion to have privacy among the communicating devices and the agents. Hence our middleware uses a decentralized architecture which supports asynchronous communication, context-awareness, context-sensitive communication, Privacy sensitive, adaptive to context-changes and reactive to emergency situation.