On detection of changes in sensor data streams
Dang-Hoan Tran, Kai-Uwe Sattler · 2011
Change detection is the process of identifying differences in the state of an object or phenomenon by observing it at different times. This problem has been researched and applied in many fields for a long time. However, designing and developing a change detection algorithm in sensor data streams is a challenging task due to the inherent difficulties in developing change detection scheme and the restricted resources of sensor network. This paper proposes a general framework for detecting changes in sensor data streams by using data synopsis structures extracted from sensor data streams. We design and implement a specific change detection algorithm by using Discrete Fourier Transform (DFT) to extract DFT coefficients as synopsis structures. We make empirical evaluations of the proposed algorithms with both synthetic and real data to show the effectiveness and the deployment potential of our change detection schemes in sensor data streams.