A Framework for Context Capturing in Mobile Learning Environments
Nowadays, learning goes beyond the achievements in educational institutions, rather it is a lifelong process. Especially, in industrial countries where knowledge has a crucial importance. That means learners are not necessarily bound to a particular curriculum, time or location. The trend goes to Ubiquitous Learning, where the central idea is the ability to learn anywhere and anytime. Since knowledge changes continually and people have to know the state-of-the-art to be able to face current challenges.
The propagation of smart phones in recent years has led to a new possibility for Ubiquitous Learning called Mobile Learning. The idea is to overcome the limitations of time and location which appear, for example, when participating in a course which takes place at a particular time and location. A new way of learning has emerged by enabling learners to learn anywhere and anytime.
The research question of this master thesis is to develop a framework that integrates different context information into one uniform model for mobile learning environments. The implemented framework should be able to collect context information about a particular user from different sources. These sources are different applications running on the mobile device monitoring the user.
All this data from different applications is sent to my framework. Here the data is collected for each user and processed. It results in a context model. A context model is the collection of all context data of a particular user in a pre-defined structure.
This resulting model can be used by other applications. Depending on an application's transformation functions they can interpret the values of the model and react early to the user by personalization, recommendation, adaptation or collecting analytics.
The necessity for this framework emerges from the obligatory need to detect and define the context of a user for every application that offers personalized services. Currently methods for capturing and handling context are strongly embedded in single applications. This circumstance limits the possibilities of context-aware applications enormously because each application captures just one part of the context. It is much more advantageous to collect context information captured by countless other applications to get a detailed overview of the user's entire context. That is why the step of context capturing and collecting has to be decoupled from single applications. By using this framework, this step is covered and does not need to be re-implemented for each application. Furthermore it is easy to integrate in existing applications.
Mobile learning applications are enriched by adding context information. In this way, a mobile learning platform can be converted to a personalized environment where the application adapts its services to the user's current context. Context describes the environment and the inner state of a user. Context-aware mobile learning applications need a detailed picture of the user's current context. Only in this way can they react immediately and precisely. This supports a better adaptation of their services and consequently it supports personalization. Detailed context capture and gathering is a complex process and therefore requires too much effort for individual applications.
In this thesis, a framework has been developed that decouples this process from the applications' logic. The implemented approach has the capability to collect context data from many and various sources and process it to a context model. Sources can be mobile learning applications or sensor information. Each application on its own can capture only a restricted part of a user's context. Only by merging this data, can one get a detailed picture.
Many similar approaches put environmental information on the same level as context. My approach is able to collect all kinds of information. It is not restricted to one type rather it is extensible. Context can describe a user's environment, a user's activities, or a user's physical state. Only such a broad knowledge can enable appropriate reactions.
Furthermore, the resulting framework is embedded in a web service that provides two interfaces, one for context sources and one for context consumers. By using the provided interfaces, developers do not have to understand the inner structure of the framework. Applications are not constrained to use a particular platform to be able to use the framework since the interaction is accomplished by HTTP communication.
Created by. Last Modification: Tuesday, 23. October 2012 07:26:05 by .