Information Discovery and Fusion: Semantics on the Battlefield 1 1 This research was supported by AFOSR contract number F49640-01-1-0542. Katia Sycara and Massimo Paolucci Robotics Institute Carnegie Mellon University Pittsburgh, PA 15213 katia@cs.cmu.edu Michael Lewis School of Information Sciences University of Pittsburgh Pittsburgh, PA 15260 ml@sis.pitt.edu Abstract To help translate information superiority to decision superiority (i.e. to rapidly arrive at better decisions than adversaries can respond to), many scientific, technical and technological challenges must be addressed. The most critical of those are information fusion and management at different levels, communication, planning and execution monitoring. Multiagent infrastructures allow information producers and users to discover one another and establish direct links. The robust, decentralized Infosphere which results, can be stood up rapidly and ensures that information of the specified types will be delivered to the right users under the right conditions. Problems involving discovery of and interoperation with new information sources are central to realizing such dynamic battlefield networks. In this paper we introduce a semantic description language and discovery technologies which may hold the keys to building large scale heterogeneous battlefield networks. 1 Introduction Today’s commanders are faced with an operational environment that is fast moving, uncertain, and flooded with information. The number of threats are bewildering and in multipolar conflicts with new and potentially shifting alliances. Therefore, just maintaining situational awareness may be a significant challenge. Add to this the new agility of rapid deployments, joint operations, and an information infrastructure that must be hastily erected in hostile environments and current C4ISR technologies are no longer adequate. Visionary concepts including Network Centric Warfare, ForceNet, the Joint Battlespace Infosphere (JBI) [11] and the Expeditionary Sensor Grid have been proposed to automate these functions but none are yet operational. The basic problems for users of current C4ISR identified by the Scientific Advisory Board on the JBI [11] were: • Information overload • Lack of interoperability • Immaturity in [higher level] fusion • Limits in display technology • Legacy tactics, techniques, and procedures Additional design problems posed in building a JBI-like infrastructure are: • Rapid network stand-up • Information source discovery • Dynamic reconfiguration Meeting these challenges requires changes in hardware, software, and the way information is represented. We have proposed [7] an agent based infrastructure to implement the publish/subscribe mechanism used by JBI to manage information. This paper deals with representational issues and illustrates how DAML-S (Darpa Agent Markup Language for Web Services), a semantic description language, could be used by agent accessible battlefield networks to address C4ISR problems. The Web Services ontology was chosen for this purpose because its tripartite distinction of profile describing a provider’s capability, model describing how the provider operates, and grounding describing how to access the provider is a good fit for the types of contextual information that would be needed to integrate battlefield information from heterogeneous sources. Consider a noncombatant evacuation operation (NEO) conducted in a central Asian country, called Irat, undergoing a revolution. There are massive amounts of information available from communication intercepts, satellite imagery, weather reports, street maps, floor plans, and maps showing gas and electric lines, as well as dynamic information sources such as spotters, air dropped sensors, and surveillance from aircraft as they enter and leave the area of interest. The commander wants to find a safe route from the US embassy where the evacuees are massed to an airport where they can be flown out of the country. His problem is that the massive amounts of