A context representation and management mechanism towards ubiquitous intelligence

The future ubiquitous intelligence is expected to be highly context-aware such that it can perceive and be aware of userpsilas situation and react accordingly. In this paper, we focus on the reusability of context by the means of distinguishing representation and management from reasoning of context. As a proof of concept, we propose a context representation and management mechanism towards ubiquitous intelligence in order to facilitate context-awareness development. It is suitable for dynamic scenes and could be coupled with different types of context reasoning algorithms. We employ our solution into a ubiquitous meeting room, which employs dynamic Bayesian network for context reasoning. Experiment results show our solution has excellent performance, configurability and extensibility.

1. Introduction

The final goal of ubiquitous intelligence is to provide the right service to the right object in the right place on the right time, which is expected to be highly context-aware such that it can perceive and be aware of user's situation and react accordingly [1]. By context, we mean a dynamic structure of information that is used to characterize the situation, viewed over a period of time, episode of use, social interaction, internal goals and local influence. In the real world, understanding of user's behavior is affected by his relations with others and the surrounding environment, therefore contextual information has to be considered [2]. Moreover, with the development of multimedia sensing technology, more and more data can be acquired by much less effort than even before. Facing large amount of available data, the use of context can help decrease the complexity significantly by processing only relevant scenarios and heuristics to choose only the objects that may involve in a particular activity.