Sensor-to-Simulation via Geoprocessing Services

This project will attempt to bridge the gap between point-based sensor observations (individual level – mobile and stationary, or samplings of fields) and the near real time provision of monitoring data, scenarios and inputs to downstream simulation and decision support environments. Bridging this gap requires the use of adequate process models, which in some cases will have to be self-calibrating, possibly with adaptive rule sets. The temporal dynamics will be primarily driven by changes in the observed values and/or locations of observations. This work will be oriented towards the domains of environmental or health modelling, mobility monitoring or risk exposure analysis.

Conceptually, research will build upon a generic model of point-to-area/field/network conversions including (but not limited to) density surfaces, semivariography, interpolation and geographically weighted regression with particular emphasis on providing input to workflows requiring field data from either sparsely sampled phenomena or from densities / frequencies. For spatiotemporal interpolation and for forward looking scenarios, individual-based modelling techniques potentially will be integrated into the overall framework.

This general framework will be built upon open services-based specifications, explicit geospatial databases, enabling distributed system and modelling architectures (sensor web and web processing specifications). As point observations typically provide (near) real time data and are mostly relevant for short cycle decision support, spanning the distance from the field to monitoring / decision points is of particular importance.

Research questions include:

• Are there any (and if so, which?) fundamental obstacles to fully integrating individual-based and aggregate simulation models into a GI-enabled environment?

• What are the most suitable (OpenGIS conformant) generic data and process models facilitating a modelling-supported near real time workflow from points to fields or areas or networks?