| Abstract: |
Today setting up a WSN deployment is a fairly straightforward engineering task. However, deployment maintenance, data quality guarantees and data interpretation often need considerable money
and labor investments. In this talk, we will shed light on the questions of automatic deployment management, data quality and data usability. We perform our analysis based on the city-scale
OpenSense air quality monitoring network operating on top of public streetcars in Zurich, Switzerland. In particular, we focus on minimizing network maintenance overhead and on improving data
quality by leveraging rendezvous of mobile sensors in the network. We explore the trade-off between locality of a process and the amount of time needed to calibrate all nodes in the network and
to check their correct functionality. Calibrated data with sufficient quality are used to build high resolution air pollution maps. Up-to-date information about urban air pollution, such as
reliable pollution maps, is of great importance for health protection agencies to timely assess the air quality situation and provide advice to the general public. The OpenSense monitoring
network was able to gather one of the largest spatially resolved and public air pollution data set available today. We develop land-use regression models to create pollution maps with a high
spatial resolution, compare the pollution maps across various time scales and study their temporal resolution limits. |
