Spatial query processing for wireless sensor networks

Abstract

Wireless sensor networks (WSN) are particularly useful for obtaining data concerning events limited to a well-defined geographic region. Applications for this task typically use spatial queries, which are SQL-like queries where location constraints are imposed on the collected data. Spatial queries save energy since only nodes inside this region collect data. This work provides a general discussion of in-network spatial query processing in WSN. We propose to divide spatial query processing in well-defined stages. This division helps understanding the spatial query processing mechanisms found in the literature and facilities to propose new solutions. This survey shows that the state of the art of spatial query processing considers regions of interest in form of rectangle or circle. To the best knowledge of the authors, no related work considers regions of interest in form of polygons, which can represent real objects plotted on maps or satellite photos. A polygon can be represented in the query by a sequence of several points, which can occupy several network packets during the query transmission, increasing the energy consumption. Furthermore, the related works, in general, consider that nodes are always on. However, nodes can go to sleep mode (turn off the radio in duty cycles) in order to save energy or can fail during their operation. This changes the network topology and, consequently, increases the amount of control message to upgrade routing tables. This work proposes a new energy-efficient fault-tolerant in-network spatial query processing mechanism that assumes regions of interest having polygon shapes, fail prone-nodes and nodes employing duty cycles. Exhaustive experiments show that the proposed mechanism process more queries than the state of the art and consume four times less energy.