earticle

영어

Recently, research on techniques of processing of KNN (K Nearest Neighbor) queries in order to find the nearest neighbor in wireless sensor networks is actively in progress. The existing representative techniques of processing of KNN queries suggest the structure-based routing technique and the non-structure-based routing technique. However, the existing representative techniques of processing of KNN queries have problems of the consumption of high energy by sensor nodes or much time spending in query processing. This paper comes up with QKNN (Quad-tree based KNN) in order to solve such problems with existing KNN query processing techniques and more efficiently process KNN queries. QKNN searches for the sensor node nearest to the query by the means of GPSR, and then composes an R-tree in order to set up the KNN Boundary with the searched node as the reference. Then, based on the R-tree, it performs parallel queries by structuring the query area in a Quad-tree in accordance with the distribution of nodes, and in each cell within the Quad-tree it processes queries by means of the non-structure based itinerary routing technique. Lastly, via various experiments using sensor data, this paper proves the excellence of the proposed technique of processing of the nearest neighbor queries.