| Summary: | 碩士 === 國立臺灣科技大學 === 資訊工程系 === 100 === Top-k monitoring facilitates the selection of the highest k numbers of sensor readings from the serial feedbacks of the nodes, and is widely utilized in distributed network applications. The major problem to the practice of top-k monitoring query is the limited energy supply of the sensors; therefore, aiming to retrench energy consumption of the wireless sensor network using top-k monitoring, we bring forth in this paper a filter-based algorithm named as Independent Adaptive Filter-based Monitoring to lessen the transmission of unnecessary messages in every sensor node. Thereby, considering the long noteless issue of probable regularity in the reading’s variation, a filter-setting step being introduced, we use probability techniques and Gaussian distribution to set an adaptive filter to each node, which automatically
adjust the future setting of filters according to the past monitoring data, to reduce the probability of the new readings going beyond the filter; moreover, we devise parameter transmission between base station and each sensor node to assure non-overlapped filter and the least feedbacks. Our proposed new algorithm is examined with both virtual and real datasets, and the simulation results prove that the new algorithm effectively reduces the energy consumption so as to considerably extend the networks lifetime compared with the previous top-k algorithms.
|
|---|
