Research And Design Of Distributed Data Query Processing Algorithms For Wireless Sensor Networks

It is a common work manner for Wireless Sensor Networks (WSNs) that, first of all, having lots of sensors distributed in a certain zone supervised, then, routing the meta data to the base station with certain routing protocol. Then users can retrieve data they need. So, it is another hot research point of WSNs to offer an efficient data query processing system for users.Aspects related with query processing were analyzed in this paper. They were data storage technology, centralize query processing algorithm and decentralized query processing algorithm and data fusion technology. Among these, centralized query processing and decentralize processing algorithm was given special study and analysis. The main idea of decentralized query processing algorithm is that, divide nodes into several zones, then parse and optimize queries at cluster headers. Cluster headers may aggregate metadata in the course of collection. Aggregated data is transported to base station at last. Two specific decentralized algorithms—Distributed Query Processing Algorithm based on Spatial Distance (DQPA-SP) and Distributed Query Processing Algorithm based on Semantic Similarity (DQPA-SS) were designed. Then a hybrid decentralize algorithm—Distributed Query Processing Algorithm based on Hybrid Model (DQPA-HM) for Heterogeneous Wireless Sensor Networks (HWSNs) was designed based the two decentralized algorithms above. Two aspects affected the query processing system were considered in this algorithm. One was the radius the cluster header can cover. And the other was the semantic similarity of metadata. With the consideration of semantic similarity, a cluster header can collect metadata with the same type. It facilitates queries'optimization and aggregation at clustering headers. It can control the zone size and control the energy consumption in zone considering the radius of cluster headers.Heterogeneous Wireless Sensor Networks (HWSNs) with the characteristic of containing kinds of sensors collect kinds of metadata. So DQPA-HM is a good choice for HWSNs's query processing because it considers the semantic similarity. This was discussed in the paper.In order to select the two parameters—the radius and semantic similarity, related experiments were conducted. Then experiments were conducted to compare DQPA-HM with centralized algorithm, DQPA-SP and DQPA-SS. Results show DQPA-HM has a better performance.