An efficient data acquisition model for urban sensor networks

Applications for Wireless sensor networks (WSN) usually take into consideration the specificity of the environment in which they are deployed in order to save the sensors' limited resources. In particular, the sensing task in urban environments requires hundreds and even thousands of sensors to be spread over the monitored area. Moreover, in environmental monitoring applications, sensors that are closely located usually provide similar readings. That is, spatial proximity is related to data similarity. In this paper we propose SIDS (Spatial Indexing Based on Data Similarity for Sensor Networks), a data model that explores this characteristic in order to provide scalability and efficient query processing on urban WSNs. Scalability is achieved by grouping sensors with similar readings, while efficiency for processing queries relies on two strategies: the concept of repositories, which consist of sensors that act as datacenters, and an indexing structure designed for speeding up both spatial and value-based queries. We have implemented the proposed model and results from simulations on a variety of scenarios show that SIDS provides scalability and it outperforms CAG and Peer-tree, which are models that have been proposed for processing data and spatial queries, respectively.