Event Detection in sensor Networks

Data generated from sensors are often expressed in time series. Monitoring and surveillance based on sensor data are important in applications of wireless sensor networks. Often it is required to specify when an event occurs so that the corresponding system changed/is changing its working status. The aim of this doctoral thesis is to explore new methods to detect events in the networks. We consider a statistical test based approach for detecting occurrences of events associated with multivariate time series. This approach allows us to automatically control the number of false alarms. We also develop a Bayesian regression method for event detection. This approach allows for different variances of noise that may exist before and after the events. For demonstration, experiments on both synthetic and real data have been conducted to show the effectiveness of the proposed detection procedures.

Author

• Tang, Debra Chiou-Shing Tzeng

Language

• en

Keyword

• Localized Event
• Event Detection
• Statistical Distribution Test
• Bayesian Approach
• Wireless Sensor Network
• Frequentist

Date created

• 2009

Type of Work

• Dissertation

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Computer Science

Degree

• Ph.D.

Advisor

• Cheng, Xiuzhen

Committee Member(s)

• Choi, Hyeong-Ah
• Bellaachia, Abdelghani
• Cheng, Xiuzhen
• Zhang, Nan
• Van De Wal, Brian James
• Rotenstreich, Shmuel

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/cr56n106c