无线传感器网络基于信号相关的目标计数

Signal Correlation Based Target Counting in Wireless Sensor Networks

解决无线传感器网络中的多目标监视与跟踪问题首先需要估计目标的个数。利用节点监听到的信号之间的相关性解决这一目标计数问题。采用的方法如下:首先利用节点检测到的信号强度对节点进行粗略分簇,使每个簇对应一个或多个相距较近的目标;然后在簇内计算不同节点采集到的信号的相关度,将信号相关度较大的节点划分为一组并对应一个目标,从而以节点组的数目估计目标个数。随后,在Matlab环境中做了仿真实验,观察了不同因素对算法性能的影响。实验表明本算法在相同条件下与已有同类算法有相当的性能,但因不要求目标信号强度相同而具有更广泛的适用性。

Target counting is an essential problem in WSN - based target monitoring and tracking application. This paper proposes a target counting algorithm based on the correlation of target acoustic signals sensed by the nodes. Sensor nodes are firstly clustered according to the sensed signal strength so that each cluster covers one or more close - by targets ; then nodes in the same cluster compute the correlation of signals and group themselves in line with the rule that nodes with higher signal correction are put into the same group, thus each group corresponds to a single target. Simulation shows that this algorithm has comparable performance to existing algorithm, however is more applicable since it relaxes the requirement that each target has the same signal strength.