I. Nedji, H. Seridi

University of Annaba, LABGED Laboratory (ALGERIA)
The joint evolution of the mobile terminals (Smartphone, PDA), mobile networks (GSM, 3G+, wireless networks, Bluetooth, etc.) and information technologies allowed to create new dimensions in computer domain. Indeed, these technologies influenced the changes of behaviors and habits of users in many sectors as work, daily life, and learning. The learning begins to take out classrooms and goes in the less classical environments, related to the learner's context. The learning becomes thus situated, contextual, personal and throughout the life. Use mobile devices to learn, anywhere and anytime, is the objective of mobile and pervasif learning. For a pervasif learning system, the ultimate goal is always to give to learner suitable learning resources according to the context. It is thus necessary to determine according to context which resources delivered, how and on which interface. All the learning process must adapt to these context which activity is it main element that gives intention and sense to different contexts and allows to define their relevant elements. In this article, we will present our pervasive learning system focusing on the activity dedicated to human learning in a mobile environment. It takes into account learners’ context to select activities and consequently select relevant learning resources according to learning context as well as learner needs and it ensure continuity of learning across the contexts. The main results obtained during the designing of our system are:

(i) a contextual scenario model for a pervasive learning that describes a common structure for learning activities highlighted across contexts,
(ii) a dynamic context management strategy based on an iterative process that uses contextual information available in the learner context and acquired from both sensors (hard and soft) and user interactions. Our context management process considers the dynamic changes of context at running moment. For this, it detects the changes in the current context then generates a new context or updates the current one in order to maintain the relevance of activity, and
(iii) a learning scenarios adaption strategy that is a set of rules basing on the contextual scenario model and the context model. To overcome the problem of the dynamic and unpredictable context, our adaption strategy is based on the combination of two adaption methods: by contract and by reflexivity. The idea is to use metadata of context properties (adaption by reflexivity) and applicable rules in certain situations (adaption by contract) in order to manage the dynamic and unpredictable context.