Adaptive Content Delivery Based on Contextual and Situational Model
Xinyou Zhao · 2010
Much of what has been written about teachers conducting field trips (ubiquitouslearning) has been anecdotal. But at most times, the learning processes have to beinterrupted during field trips because of lack of support from school administrationsand inadequacy of resources.Although e-learning system may share the learning objects with ubiquitouslearning, most of learning objects designed for desktop computers and high-speednetwork connections today, are not suitable for network features with low bandwidthand handheld devices with limited resources and computing capabilities. So up tonow, access to e-learning objects by ubiquitous devices has not become asconvenient as expected.The objective of this research is to provide adaptive content delivery in a commondynamic and extensible framework in order to have scalable, flexible and extensiblearchitecture, which can dynamically define an adaptation scheme thatrecodes/reconstructs learning objects according to context awareness of learners,such as context data, learning object, and learner’s preference. The proposal isconsisted of three services to improve the usability of e-Learning objects onubiquitous device: 1, content service; 2, transcoding service; 3, presentation service.The research in this thesis has concentrated on the issues of designing andimplementing learning system that should meet the learning context awareness inubiquitous learning environment. In this research, we propose a new method basedon parallel production system (two working memories: context and situation), whichdynamically transcodes the e-learning objects according to learning context andsituation. Context method determines that the learning preference (preferred mediatype) by multiple-response reward-penalize algorithm and situation method providesIVadaptive contents according to learning context (device, network, and preferredmedia type) by negotiation algorithm. Interpreter determines the final version ofaccessed contents by Rete algorithms. We call this process as content service.The transcoding service analyzes the original learning objects (not writtenspecifically for mobile phones and devices and thus not displayed properly or stored)and converts it into a mobile-ready format. The proposed caching mechanism basedon proposed feature finite state machine (FFSM) in transcoding service reduces thetranscoding time.At last, the presentation service pushes recoded contents to remote learners,which contents are adaptive to learning context and situation.Numerical examples (learning Chinese course and standard document onubiquitous device) are given, and properties of the methods are examined at last. Theexperiment results show that the adaptation process satisfies the request of system(e.g. low computing resource, quick response, etc). Because the proposal can providea seamless learning with e-learning system at any time and place with any device, theresults show that the learning experience is improved in ubiquitous learningenvironment by this research.