Planning under pressure in time-constrained environments while relying on uncertain information is a challenging task. This is particularly true for planning the response during an ongoing disaster in a urban area, be that a natural one, or a deliberate attack on the civilian population. As the various activities pertaining to the emergency response need to be coordinated in response to multiple reports from the disaster site, a user finds itself cognitively overloaded. To address this issue, we designed the Anytime Cognition (ANTICO) concept to assist human users working in time-constrained environments by maintaining a manageable level of cognitive workload over time. Based on the ANTICO concept, we develop an agent framework for proactively managing a user’s changing information requirements by integrating information management techniques with probabilistic plan recognition. In this paper, we describe a prototype emergency response application in the context of a subset of the attacks devised by the American Department of Homeland Security.
|Title of host publication
|The Fifth Annual Conference of the International Technology Alliance
|University of Pittsburgh Press
|Published - 2011