基于多模型预测的WSN动态簇目标跟踪算法

分析了无线传感器网络(WSN)中现有目标定位算法缺陷,针对WSN中跟踪目标的运动不确定特性,提出一种基于多模型动态簇预测的WSN目标跟踪算法;该算法以多模型为目标动态建模框架,适应目标的不确定运动,在得到各模型预测之后,综合各模型预测估计形成全局预测估计,进而构造局部唤醒区域,将落入该局部唤醒区域的节点构造动态簇,通过择优规则,确定该节点动态族的簇首,实现目标的定位跟踪;仿真结果表明,与全局唤醒算法相比,所提出的基于多模型动态簇预测的WSN目标跟踪算法可适用于目标的不确定运动,并得到较好的跟踪精度。

A Dynamic Cluster Target Tracking Algorithm for WSN Based on Multi-model Forecast

Disadvantage of the target localization algorithm in WSN(Wireless Sensor Network) has been analyzed and a target tracking algorithm for WSN based on dynamic cluster forecast by multi-model has been proposed regarding the uncertainty of the target's movement in WSN.In order for better adaptability into the case of the uncertain movement of the target,the proposed algorithm uses multi-model to construct the dynamic modeling frame for the target.The dynamic cluster can be determined by the global forecast of the multi-model and the cluster head can be chosen by a rule.The simulation result shows that our algorithm has higher tracking precision and better adaptability in the case of uncertain movement of the target,compared with the global waking algorithm.