A Process Chain for End-to-End Sensing in Disruptive Environments

Abstract

Sensor networks were conceived as a novel tool for distrubuted sensing, e.g. in environmental sciences. Initially it was thought that inherent imperfections in the data would be "oversampled", however this is only theory (so far). As a single sensing point in space and time is still expensive, many applications require accuracy in absolute terms and users are not felxible (yet) to deal with lossy data types this assumption has to be reconsidered. We argue that the high up-front investments of WSNs require a reliable interaction of all system components and layers and that for the user an exact knowledge of the origin nature of data along with the whole processing chain is key. We will introduce concepts and demonstrate initial solutions in the context of the PermaSense architecture used for high alpine environmental monitoring in the Swiss Alps.