iPois is an Intelligent Traveler Information System based on a generic selective attention-based, multi-agent information system. The architecture of the system involves a master agent and personal agents for each registered traveler.
The main role of the master is collecting selective information from several web location-based services, and send it to the personal agents so that they can selectively deliver information to the several mobile devices, owned by humans. Each personal agent models a user and acts on his/her behalf. For this reason it comprises the following models: a model for beliefs/expectations, i.e., a model for representing and generating expectations (each one of the personal agents has information about the expectations of its owner, based on their travel history); a model for representing desires and their dynamics; a model for feelings; and a model for selective attention. It is worth noticing that the master also exhibits the same models, with the exception of the model for representing and generating desires.
The client application (iPois) comprises an interface so that users can interact with features of their personal agents. It allows: the visualisation of the agents' expectations for all the POIs, given the user's context (different periods of the day, day of the week and goal); the configuration of the visual information in the screen, i.e., by turning on or off the selective attention filters (surprise and uncertainty), their spaciotemporal coordinates, radius and goal; to semi-automatically learn from an experience, by allowing the user to give feedback to the agent. As each user has unique preferences, intentions and behaviors, the personal agents will have different expectations for different users, even if their users' context is the same.