k Nearest Neighbor Search for Location Dependent Sensor Data in MANETs in NS2

k Nearest Neighbor Search for Location Dependent Sensor Data in MANETs in NS2

Abstract:

K nearest neighbor (kNN) queries, which retrieve the k nearest sensor data items associated with a location (location-dependent sensor data) from the location of the query issuer, are useful for location-based services in mobile environments. Here, we focus on the kNN query processing in mobile ad hoc networks (MANETs). Key challenges in designing system protocols for the MANETs include low-overhead adaptability to network topology changes due to node mobility, and query processing that achieves high accuracy of the query result without a centralized server. In this paper, we propose the filling area (FA) method to efficiently process kNN queries in the MANETs. The FA method achieves low overhead in query processing by reducing a search area. In the FA method, data items remain at nodes near the locations with which the items are associated, and nodes cache data items whose locations are near their own so that the query issuer retrieves kNNs from nearby nodes. Through extensive simulations, we verify that our proposed approach achieves low overhead and high accuracy of the query result.