1. Introduction

The emerging mobile Internet makes it possible for users to access a myriad of services while on the move. At the same time, it imposes constraints that require additional levels of automation:

§ Mobile devices tend to have very limited input/output functionality

§ Mobile users often face time-critical, goal-driven tasks (e.g. looking for a nearby gas station in the case of a civilian user or for battleground status updates in the case of military users). Often they are also subject to many more distractions than desktop users (e.g. driving through busy intersections, talking to colleagues, or possibly facing enemy fire).

The Semantic Web offers the prospect of matching user contexts with relevant services, while minimizing the amount of information required from the user.

2. Objectives

Our project aims at extending existing Semantic Web tools and ontologies in support of mobile context-aware scenarios. Target applications include both:

§ Civilian mobile Internet scenarios: With mobile phone ownership set to reach the billion mark in 2002 and with already tens of millions of Internet-enabled mobile phones worldwide, the mobile Internet and associated mobile commerce scenarios offer a unique opportunity for the broad adoption of Semantic Web technologies

§ Military mobile context-aware scenarios: We plan to extend our technology and demonstrate context-aware alert filtering functionality in support of DoD scenarios to be identified jointly with other members of the DAML initiative and prospective DoD users in the context of the 2002 DAML Experiment.

3. Work

Initial work will be carried out in the context of a campus-based Semantic Web demonstrator that leverages the IEEE802.11 Wireless LAN (WLAN) at Carnegie Mellon University (CMU). With its coverage, which spans the whole campus, and its support for location tracking functionality, CMU’s WLAN provides an ideal environment for rapid prototyping and early evaluation of mobile context-aware functionality.

Work will include development of a set of DAML+OIL ontologies, DAML-S service descriptions and rule-based preferences and permission profiles (embedded in specialized agents) to support a collection of mobile context-aware services. Agents to be developed will demonstrate both push and pull functionality. An example of a pull functionality could be in the form of a restaurant concierge that advises members of the campus community on places where to have lunch, taking into account their current position, as well as their calendar activities (e.g. when their next class is scheduled) and the weather (e.g. is it sunny or raining?). Push functionality will involve the development of customizable, rule-based filtering agents that select which alerts to display to users depending on their current context (e.g. “when in class, do not disrupt me with promotional messages”, or “when it is lunch time, show me only promotional messages from nearby restaurants”). The resulting functionality will be tested with members of the campus community who will be connecting to the Internet from PDAs over the wireless LAN.

In the process, our project will propose extensions to DAML+OIL and DAML-S aimed at supporting mobile service descriptions and at capturing contextual attributes and context-sensitive preferences. DAML+OIL ontologies and DAML-S service descriptions developed during the course of the project will be made available to the DAML community and results will be demonstrated to prospective DoD and civilian customers. In addition, we propose to extend our context-aware message filtering functionality in support of scenarios to be identified in the context of the 2002 DAML Experiment.