Title of Presentation: Decision Task Driven Dynamic Configuration of Virtual Sensorweb

Primary (Corresponding) Author: Prasanta Bose / Ryan Timmons

Organization of Primary Author: Lockheed Martin Space Systems, ATC

Co-Authors: Ryan Timmons

Abstract: A key challenge in weather/science analysis and prediction tasks that involve sensing, monitoring and aggregation of events across time, space and sensing domains is dynamic configuration and robust coordination of distributed information sources driven by semantic decision models and user demands. The information source nodes in such a sensorweb architecture correspond to gateways to physical sensors and sensor networks. The NASA AIST funded VSICS project is developing a middleware infrastructure that enables flexible and robust management (binding of services to requests) and distributed coordination of such gateway resources (called virtual sensors in VSICS) that maximize user utilities. This paper presents initial results on implementing such an architecture and applying to prediction of lightning events (a primary cause of wildfires) and exploitation of such events to semantically mark cloud imageries obtained from continuous satellite based cloud-imaging devices for focusing analysis and prediction.