Processing k nearest neighbor queries in location-aware sensor networks

Efficient search for k nearest neighbors to a given location point (called a KNN query) is an important problem arising in a variety of sensor network applications. In this paper, we investigate in-network query processing strategies under a KNN query processing framework in location-aware wireless sensor networks. A set of algorithms, namely the geo-routing tree, the KNN boundary tree and the itinerary-based KNN algorithms, are designed in accordance with the global infrastructure-based, local infrastructure-based and infrastructure-free strategies, respectively. They have distinctive performance characteristics and are desirable under different contexts. We evaluate the performance of these algorithms under several sensor network scenarios and application requirements, and identify the conditions under which the various approaches are preferable.