Improved Multi-Dimensional Top-K Query Processing Based on Data Prediction in Wireless Sensor Networks

Abstract

Since the scale of wireless sensor networks is expanding and one single node can sense a variety of data, selecting the data of interest to users from a tremendous data stream has become an important topic. With further development in the field of WSN query, extensive research is being conducted to solve different kinds of query issues. Skyline is a typical query for multi-criteria decision making, and many applications have been developed for it. Studies of multi-dimensional top-k query processing have proven it to be more efficient than traditional centralized scheme. In some cases, variations of observed conditions, such as temperature and humidity, are related to time. Thus, we used a data-prediction method to establish the bi-boundary filter rule, which helps filter the data that may be dropped by the sets of final result. The bi-boundary filter rule determines whether the received or generated data will be transmitted. We analyzed the simulation results and concluded that the bi-boundary filter rule can be more energy-efficient in situations in which a temporal correlation exists.