量化量测条件下的交互多模型箱粒子滤波

发布时间：2018-10-23 16:39

【摘要】：在分布式多传感器网络中,为了节省通信带宽,需要将传感器得到的点量测量化成区间量测,而传统的滤波算法均不能直接处理这种量化量测.箱粒子滤波作为一种"广义粒子滤波"算法,用箱粒子和误差界限模型来取代传统的点粒子和误差统计模型,是新近出现的处理区间量测的有力工具.相比粒子滤波,箱粒子滤波还具有所需粒子数少、算法复杂度低、运行速度快等优点.因此,为了处理量化量测条件下的机动目标跟踪问题,提出了交互多模型箱粒子滤波算法.仿真对比实验表明:在量化量测条件下,交互多模型箱粒子滤波算法和交互多模型粒子滤波算法都能够准确地估计机动目标状态,但交互多模型箱粒子滤波所需粒子数更少、计算效率更高.
[Abstract]:In distributed multisensor networks, in order to save communication bandwidth, it is necessary to transform the sensor's point measurements into interval measurements, which can not be directly processed by traditional filtering algorithms. As a "generalized particle filter" algorithm, the box particle filter replaces the traditional point particle and error statistical model with the box particle and error bound model. It is a new powerful tool for dealing with interval measurement. Compared with particle filter, box particle filter also has the advantages of small number of particles, low algorithm complexity and fast running speed. Therefore, in order to deal with the problem of maneuvering target tracking under the condition of quantitative measurement, an interactive multi-model box particle filter algorithm is proposed. The simulation results show that both the interactive multi-model box particle filter algorithm and the interactive multi-model particle filter algorithm can accurately estimate the state of maneuvering target under the condition of quantitative measurement, but the particle number of interactive multi-model box particle filter is less. The calculation efficiency is higher.
【作者单位】：西安电子科技大学电子工程学院;
【基金】：国家自然科学基金资助项目(61372003) 中央高校基本科研业务费专项资金资助项目(JB140221)
【分类号】：TN953

【共引文献】