Today, the internet gives access to huge amounts of time- and location-related information, such as local events, time tables of various means of transportation, or news. Additionally, social network sites, such as Facebook or Twitter add even more information from friends and peers tailored to certain social groups sharing common interests. The vast adoption of mobile phones and smart phones provides a widely deployed gateway to this cloud of information while being on the move anytime anywhere. As the attention of mobile users is always split among many simultaneous tasks, valuable location-based services have to be tailored to the users’ current interests and needs.

In order to implement such a solution, several research areas must be brought together.

1. • Location based services are becoming accessible through mobile devices via the mobile network and make use of the geographical position of the mobile device in order to adapt their responses to the current situation of the user.
   

2. • Pervasive and ubiquitous computing aims at developing new models of human-computer interaction that thoroughly integrate information processing into everyday objects and activities that may be located anywhere. “Intelligent” applications provide location, situation and user adaptive information and in this way greatly improve the ability of information systems to successfully assist users.
   

3. • Mobile recommendation systems add another useful capability: they support a mobile user or a group of mobile users when making decisions ’on the go’, whereby they may leverage shared human experience, the ‘wisdom of the crowds’.
   

On the very bottom location data and context derived from various sensors can act as an indicator for the user’s current activities. By complementing the sensor data with domain knowledge of a specific assistance scenario (like shopping, or health-care) we can go well beyond this functionality: In assistance scenarios – in contrast to a general setting – user activities can be assumed to be related to certain well defined tasks. By formalizing those tasks and the related user goals, additional meaning can be attributed to context information. On this basis the integration of user preferences can be achieved, keeping the user in the loop of control during concurring and uncertain system states.

We shall emphasize that the fields and applications mentioned above share similar problems, currently dealt with separately in the distinct scientific communities.