An Aggregation Tree Approach for Event Detection in Wireless Sensor Networks

Abstract

Collecting data with confidence of the environment is an important application of the wireless sensor networks (WSN). In this study, we used a splay tree based event region fault-tolerant detection algorithm (STERD) for WSNs to address these problems. Rather than passing large quantities of raw data through the network, our network instead sends the coefficients of a regression polynomial that simulates the readings of the sensor nodes throughout the region of interest. These coefficients essentially summarize the relevant information over the region, in effect compressing the data, leading to a reduction in both data volume and energy expenditure. Simulation results indicate that the proposed STERD can attractively obtain the high event region detection accuracy and considerably reduce the false alarm probability.