Context Management for Mobile Devices

New and rich environments for ubiquitous computing require the development of open, scalable and flexible solutions for multiple problems, including context management. Among these, the necessity to evolve the notion of context is being seen as one of the most crucial.

In this work it is expected an innovative representation for the location context of a mobile user by aggregating contextual information in a multidimensional space. The approach is based on an open and generic entity capable of supporting virtually any sensor or location service without impose any specific space model. Moreover, it is capable of supporting data globally useful, at the same time that holds data linked to local or specific space models. The solution, with a set of normalized fields, allows the development of applications without being dependent on the sensors used to acquire the contextual data. The generic representation supports the addition of new data without imposing any space model and leads to a generic solution capable of dealing with new sensors and contextual data. Supporting any space models, the architecture grants the existence of contextual data with meaning only to a specific place as it supports global data. We are now further developing the context representation by adding to the Context Manager the capability to add calculated and estimated attributes. Considering that most of people in European countries have a GSM mobile phone, we are developing a mathematical model that, based on cell identification, without knowing the network configuration or demanding any user intervention, identify places visited by the user and provide a user familiarity level for each one. This computation is possible to be done using only cell-id information available on the mobile phone, without using any network service.