摘要
为了实现对无线传感器网络中移动目标的分布式跟踪,提出了一种基于二阶段的完全分布式迭代目标跟踪算法。算法的每次迭代都由2个阶段构成,即估计阶段和达成一致性阶段。在估计阶段、网络节点或者通过它的测量值来估计出目标的位置,并通过卡尔曼滤波来改进这个测量值,或者根据嵌入的卡尔曼滤波线性运动模型对目标运动进行预测;在达成一致性阶段,通过执行一种最大一致性策略,使得由单个节点单独完成的估计值收敛到一个共同的关于目标位置的最佳值。仿真结果表明,分布式迭代跟踪算法尽管只有一小部分代理可以在每个时刻感知到目标,但能够很好地预测目标的运动轨迹,并有效地提高跟踪精度。
In order to realize the distributed tracking of moving target in wireless sensor networks,a completely distributed iterative target tracking algorithm based on the two phases is proposed in this paper. Each iteration of the algorithm consists of two phases,namely an estimation phase and a consensus phase. In the estimation phase,the network nodes estimate the position of the target through its measurement value and improve the measurement value by a Kalman filter,otherwise will predict the target motion according to the embedded linear motion model of the Kalman filter. In the consensus phase,the estimation performed individually by single node converges on a common and best value for the target position by executing a maximum-consensus strategy. The simulation results show that the proposed distributed iterative tracking algorithm,in which although only a small part of the agents can sense the target,can predict well the target trajectory and improve the tracking accuracy effectively.
出处
《电子测量与仪器学报》
CSCD
北大核心
2018年第1期194-200,共7页
Journal of Electronic Measurement and Instrumentation
