Models of People, Places and Devices for Location-Aware Services

Posted: July 17th, 2006 | No Comments »

Assad, M. J Kay and B Kummerfeld 2006 Models of people, places and devices for location-aware services. In
proceedings of 4th International Pervasive Conference, Late Breaking Results.

The authors provide an overview of their Active Model (AM) to manage all elements of a ubiquitous environment: sensors, services, devices, rooms, and people. This contrast with work on other systems (work by Armind Dey or Jeffrey Hightower) that focus on managing data from sensors.

Their system uses an accretion/resolution representation. It provides a hierarchy of contexts, each acting as a namespace for the component modeled. AM uses subscriptons attached to components to drive information flow between models and other programs in context-aware applications. It has a service discovery that enables models of sensors, devices, people and places to be stored at arbitrary machines.

Active Model Figure1

Active Model Figure3

Relation to my thesis: a simple model for context-aware systems. I could use this model as a base to integrate the technical/physical constraints (the fuzziness of sensed space) and the layer of uncertainty inherent to “real-world” context-aware environments. Figure 2 shows a location information triggered with a user passing by proximity of a BT sensor. This can raise multiple misleading behaviors and informations (e.g. I am not at location g46, but I passed by in the hallway). The ad-hoc exchange of remote location without Internet connectivity raises the question of location timeliness.