Abstract
Context-aware computing is a class of mobile computing that can sense its physical environment and adapt its behavior accordingly; it is a component of the ubiquitous or pervasive computing environment that has become apparent with innovations and challenges. This chapter reviews the concept of context-aware computing, with focus on the user activities that benefit from context history. How user activities in the smart environment can make use of context histories in applications that apply the concept of context prediction integrated with user pro-activity is explored. A brief summary of areas which benefit from these technologies as well as corresponding issues are also investigated.
TopContext And Context-Awareness
While most people tacitly understand what context is, they sometimes find it difficult to elucidate. The term “context awareness”, was first introduced by Schilit and Theimer (Schilit and Theimer 1994). Their definition of “context” is “the locations and identities of nearby people and objects and changes to those objects”. This definition is useful for mobile computing. It defines context by examples, and thus is difficult to generalise and apply to other domains.
Winograd points out that context are composed of “con” (with) and “text”, and that context refers to the meaning that must be inferred from the adjacent text. Such meaning ranges from the references intended for indefinite articles such as “it” and “that” to the shared reference frame of ideas and objects that are suggested by a text (Winograd 2001). Context goes beyond immediate binding of articles to the establishment of a framework for communication based on shared experience. Such a shared framework provides a collection of roles and relations with which to organise meaning for the phrase.
Key Terms in this Chapter
Regularity: The probability of user mobility in following the user’s daily habits. Regularity basically monitors user mobility and follows the user’s regular movements in the Smart Office.
Context Prediction: The prediction of future context based on recorded past context i.e. context history is often conceived as the ultimate challenges in exploiting context histories.
Context: Defined as rich and rapidly changing predicate relations between objects (user and environment entity) that contain information relevant to the current local domain while an object (user entity) is on the move.
Location Estimation: The use of proximate sensor data using machine learning algorithm to estimate a user location.
Pattern Accuracy: The adjustment of the degree of the regularity of user mobility to actual user mobility. By increasing the level of pattern accuracy, the user’s mobility pattern (map) can be improved.
Context-Aware Computing: Defined as a new software engineering approach in the design and construction of a context-aware application which exploits rapid changes in access to relevant information and the availability of communication and computing resources in the mobile computing environment.
Location Prediction: The use of probabilistic method to predict user location based on patterns of historical data of fixed sensors and proximate sensors data.