Gradient Boundary Detection for Time Series Snapshot Construction in Sensor Networks

• Authors: Jie Lian;Lei Chen;Kshirasagar Naik;Yunhao Liu;Gordon B. Agnew
• Venue: IEEE Transactions on Parallel and Distributed Systems
• Year: 2007

In many applications of sensor networks, the sink needs to keep track of the history of sensed data of a monitored region for scientific analysis or supporting historical queries. We call these historical data as a time series of value distributions, or snapshots. Obviously, to build the time series snapshots by requiring all the sensors to transmit their data to the sink periodically is not energy-efficient. In this paper, we introduce the idea of gradient boundary, and propose a gradient boundary detection (GBD) algorithm to construct these time series snapshots of a monitored region. In GBD, a monitored region is partitioned into a set of sub-regions and all sensed data in one sub-region are within a predefined value range, namely gradient interval. Sensors located on the boundaries of the sub-regions are required to transmit the data to the sink, and then the sink recovers all sub-regions to construct snapshots of the monitored area. In this process, only the boundary sensors transmit their data, and therefore, energy consumption is greatly reduced. The simulation results show that GBD is able to build snapshots with a comparable accuracy and has up to 40% of energy saving compared with the existing approaches for large gradient intervals.