论文部分内容阅读
如何能量高效的进行top-k查询处理是无线传感器网络领域中的一个重要课题.节点设置过滤窗口可以避免与top-k查询无关的数据上传到汇聚节点或者基站,因而大大减少传感器网络的通信量,节省传感器节点能量.然而,已有算法如FILA、DAFM,基站到传感器节点的过滤窗口更新中仍然存在很大开销.提出一种基于预测信息更新窗口的top-k查询算法FAPU,该算法根据历史数据采用ARIMA时间序列预测模型对接下来s个时刻的传感器数据进行预测,根据预测信息进行多步窗口更新的代价评估,避免不必要的窗口更新,从而减小窗口更新的能量消耗.实验结果表明在确保top-k查询准确性的同时,本文所提出的FAPU算法与已有算法相比更加能量有效.
How energy-efficient top-k query processing is an important issue in the field of wireless sensor networks Node set up a filter window to avoid the top-k query has nothing to do with the data uploaded to the sink node or base station, thus greatly reducing the sensor network traffic , Saving sensor node energy.However, there are still a lot of overhead in existing filtering algorithms such as FILA, DAFM and base station to sensor node update.This paper proposes a top-k query algorithm FAPU based on the prediction information update window, The historical data uses ARIMA time series forecasting model to predict the sensor data of the next s time points, and based on the forecasting information, the cost of updating the multi-step window is evaluated to avoid unnecessary window updating, so as to reduce the energy consumption of window updating. While ensuring the accuracy of top-k queries, the proposed FAPU algorithm is more energy-efficient than the existing algorithms.