networks play an important role in applications concerned with
environmental monitoring, disaster management, and policy making.
Effective and flexible techniques are needed to explore unusual
environmental phenomena in sensor readings that are continuously
streamed to applications.
In the first of a series of papers, we proposed a framework that allows to detect outlier sensors and to efficiently construct outlier regions from respective outlier sensors. We then presented a system, called ORDEN, that allows for the detection and (visual) exploration of outliers and anomalous events in sensor networks in real-time.
The next paper presented an approach that facilitates the efficient computation of anomaly regions from sensor readings with a particular focus on obstacles present in the sensor network. We improved this approach by proposing a distributed in-network processing technique where the region detection is performed at the sensor nodes. We demonstrated the advantages of this strategy over a centralized processing strategy by utilizing a cost model for real sensors and sensor networks.
We use real-world sensor data from different application domains to evaluate our techniques. Our three main data sources are:
- CIMIS: sensor data from weather sensors in California
- TAO: sensor data from moored ocean buoys in the Pacific Ocean
- Intel Lab: data from 54 indoor sensors
We store all of the Intel lab data and part of the CIMIS and TAO data in a local Postgres database for easy access.
We implemented a prototype for all outlier region detection tasks. The code is written in Python. It uses Tkinter for the GUI and 2d visualizations, and VTK for the 3d visualization. Please contact Conny Junghans if you are interested in the code.
In-Network Detection of Anomaly Regions in Sensor Networks with Obstacles. Conny Franke, Marcel Kernstedt, Daniel Klan, Michael Gertz, Kai-Uwe Sattler, Elena Chervakova. In Computer Science - Research and Development (Springer), 2009.
ORDEN: Outlier Region Detection and Exploration in Sensor Networks (demo paper). Conny Franke and Michael Gertz. In ACM SIGMOD International Conference on Management of Data, SIGMOD 2009.
In-Network Detection of Anomaly Regions in Sensor Networks with Obstacles. Conny Franke, Marcel Kernstedt, Daniel Klan, Michael Gertz, Kai-Uwe Sattler, Elena Chervakova. In 13. GI-Fachtagung Datenbanksysteme für Business, Technologie und Web, March 2009, Muenster, Germany, pages 367-386, 2009.
Detection and Exploration of Outlier Regions in Sensor Data Streams. Conny Franke and Michael Gertz. In Workshops Proceedings of the 8th IEEE International Conference on Data Mining (ICDM 2008), December 15-19, 2008, Pisa, Italy. IEEE Computer Society 2008, pages 375-384, 2008.