基于伸展树的无线传感器网络事件区域检测

An Event Region Detection Scheme Based on Splay Tree in Wireless Sensor Networks

本文提出一种基于伸展树的事件区域容错检测算法,该算法在构成的分布式融合树上实施,由多元回归方程对事件发生的可能性进行评估,利用被检测事件的时空相关性,使得在缺少传感节点的区域,Sink节点也可以获得事件发生的状态信息.理论分析及仿真结果表明,即使节点发生故障的概率较高,提出的检测算法也能获得较高的精度和较小的误判率,表明在事件区域容错检测中应用该方法是可行的.

This paper investigates a kind of algorithm called splay tree based event region fault-tolerant detection algorithm （STERD） for Wireless Sensor Networks （WSNs）. Since the algorithm is implemented over a splay tree in a distributed manner, a polynomial regression addressed these problems of event region detection by having a splay tree of sensor nodes. The sink can get attribute values in regions devoid of sensor nodes for attribute values that show smooth spatio-temporal correlation. Theoretical anal- ysis and simulation results indicate that the proposed STERD can atlractively obtain the high event region detection accuracy and considerably reduce the false alarm probability even when fault nodes exist.