Article Preview
Top1. Introduction
Proliferations of mobile devices in people’s everyday life spurred by the significant advancement in mobile computing have taken personal computing beyond the desktop level. Today’s mobile devices have evolved from merely phones to full-fledged computing, sensing, and communication devices. Fueled with remarkable market demand, mobile phone providers are working around the clock on enhancing their products with innovative value-added services and rich content applications. By taking a quick glance at most existing mobile devices, one can notice a long list of value-added personal digital assistance applications, ranging from location-based information retrieval and messaging services to embedded ambient and touch sensors to detect environment and automatically adapt the applications behavior accordingly.
Amid this evolving trend, context-aware computing, which refers to computational systems that can sense clues about the users’ context and enable desired interaction between those systems and users, has gained increasing recognition as one of the emerging technologies for the next generation of mobile devices. Such context-aware systems typically acquire the context using sensors or user behaviors to perceive a situation, and then try to match the perceived stimulus to context in order to adapt the application based on the recognized context.
Today’s smart phones are typically equipped with various built in sensors to measure motion, environmental conditions, and location. Motion sensors measure acceleration forces and rotational forces along three axes; environmental sensors measure various environmental parameters, such as ambient air temperature and pressure, illumination, and humidity; while position sensors measure the physical position of a device. These sensors are capable of providing high precision, accurate raw data that can be used to make inference about user personal context and behavior. In addition to sensor-based data, user activities such as searching the web or checking emails may also provide more clues about user's personal context.
Mobile devices use the collected context to smartly recognize and interpret their surrounding environment and then react accordingly. This is typically done by context-aware applications that adapt their interfaces, tailor the set of application-relevant data, increase the precision of information retrieval, discover services, make the user interaction implicit, or build smart environments (Bolchini et al., 2007). For example, a context-aware mobile phone may sense nearby Bluetooth-enabled mobile devices within its range and enable users to exchange and share information to facilitate a social interaction (Persson & Jung, 2005). Another example of a context-aware application is a location-aware tour guide that provides real-time attraction guidance and points of interest based on users’ current location and set preferences (Schwinger et al., 2007).
Much work has been devoted on sensing and inferring context information, yet the relevancy among these multifaceted contexts is rarely investigated. The main challenge that renders context-aware services lies not so much in acquiring data, but in determining which collection of data best represents the current context (Anderson et al., 2006). While sensors can gather information relevant to context, such as location, acceleration, light levels, orientation, etc., the connection between the collected context data and the usage of these data in mobility has not been investigated (Barnard et al., 2007). This results in a deficit of understanding of the entangled relationship among the context data in reality.
In this paper, we propose a framework for modeling and evaluating collected context data and the nature of relevancy among the data. The framework examines a comprehensive and integrated context awareness approach that aims at linking context awareness to multiple context data, and quantifying the interrelationship among these data. Through the estimation of such a model, it becomes possible to isolate the contribution of each context data towards context awareness. The framework provides preliminary guidelines for determining representative measures of context data in the design and evaluation policies of context awareness in mobile human computer interaction.