Microsoft Research Faculty Summit 2008
Liqian Luo Networked Embedded Computing Microsoft Research
SenseWeb is a peer-produced sensor network to enable sharing of sensors and data across the globe Sensor registration and metadata management Data publication and global sharing Spatio-temporal data exploration SwissEx: weather stations Taiwan: rain sensors Citysense: urban sensor network Asia-Pacific Environmental Sensor Grid
Heterogeneity Sensor hardware Resource capability: bandwidth, power, computation Willingness to share Measurement accuracy Scalability Streaming all raw data from all sensors to all applications not feasible Security and Privacy Data Verifiability, Trust
User contributed sensors and data Each scientist deploys at small scale; everyone can use shared instrumentation Larger spatio-temporal coverage than any single system Costs amortized over multiple experiments Same datasets used for multiple analyses Features: Flexibility in creating new sensor types/schemers Flexibility in registering sensors and data sources
Real-time data presentation Enable real-time monitoring of sensors Features: Snapshot data view to inspect the latest data entry Time series chart to observe data trend in real-time
In-situ data analysis Discover data correlations, dependencies, and distributions via one click Features: Comparison chart to view correlations of multiple sensors (such as temperature, humidity) Contour maps overlaid on top of terrain maps to view spatial distributions Visualize spatio-temporal distribution by overlaying data from different times