无线传感器网络下的并行粒子滤波目标跟踪算法

Target Tracking Algorithm Based on Parallel Particle Filtering in Wireless Sensor Networks

针对无线传感器网络环境下目标跟踪问题,提出一种基于分布式并行粒子滤波的目标跟踪方法。在建立了网络动态分簇模型和目标运动模型的基础上,将并行粒子滤波算法应用于动态目标进行跟踪。算法通过多个感知节点并行的运行局部粒子滤波器,得到每个节点对目标状态的估计,动态成簇的簇头节点对簇内每个节点的信息进行融合,形成动态目标的状态估计,提高了目标跟踪的精度。同时通过动态簇头之间的目标状态信息的交换,实现了运动目标的动态连续跟踪。仿真结果表明,算法实现了运动目标协作跟踪,与集中式结构目标跟踪相比,跟踪精度提高了30%。

Focusing on the problem of target tracking in the wireless sensors networks, a parallel collaborative tracking algorithm based on particle filter is proposed. This algorithm is applied to the target tracking based on wireless sensor network dynamic clustering model and target motion model. The tracking accuracy is improved by using the parallel particles filters collaboratively with multi-sensors nodes. The optimal target tracking is implemented by using the information fusion and exchanging target states between the dynamic clusters heads. The simulation results show the target tracking problem in the wireless sensor networks can be solved better by the proposed algorithm, the precision of tracking is increased by about 30% compared with the traditional particle filtering algorithms.