摘要
协同导航技术是提升平台协作性能的重要保障和关键技术,针对复杂环境中导航信息测量数据丢包或延迟问题,提出一种协同导航滤波用不完全量测环路和积数据关联算法(IM-LSPADA),将局部节点状态与友邻节点状态进行扩维,协同节点的状态噪声联合扩维,为系统状态变量,友邻节点测距为观测量,对状态与量测噪声的后验概率密度函数进行高斯近似;量测数据随机延迟或丢包时,采用上一时刻量测量作为系统观测值,基于确定积分点进行采样的贝叶斯框架,计算预测目标节点位置,进行定位。通过无迹变换(UT)传播的sigma积分点进行IM-LSPADA估计仿真和实验结果表明,量测数据丢失时,能够完成目标网络的定位和跟踪。与未考虑量测随机延迟的SPBP算法相比,改进算法的横轴位置误差降低了76%,纵轴位置误差降低了66%,精度可达到标准的和积数据关联算法(SPADA)的精度。
Cooperative navigation technology is an important guarantee and critical technology to improve platform collaboration performance.Aiming at the problem of packet loss or delay of navigation information measurement data in complex environment,this paper proposes a cooperative navigation filtering of incomplete measurement loop sum-product algorithm for data association(IM-LSPADA).The local node state and neighboring node state are dimension expanded,and the state noise of the cooperative node is jointly dimension expanded as the system state variable.The measured value of the distance among adjacent nodes is used as the observation value.The Gaussian approximation is performed on the posterior probability density function of the state and the measurement noise.When the measured data is randomly delayed or the data packages are lost,the previous measurement is used as the system observation value,and the predicted target node position is calculated and the positioning is performed based on the Bayesian frame of determining the integration point for sampling.The results of simulation and experiment of IM-LSPADA estimation through the sigma integration point propagated by unscented transformation(UT)show that when the measurement data are lost,the positioning and tracking of the target network can be completed.Compared with the SPBP algorithm that does not consider the measurement random delay,the improved algorithm reduces the horizontal axis position error by 76%and the vertical axis position error by 66%,and the accuracy can reach the accuracy of the standard sum-product algorithm for data association(SPADA).
作者
陈红梅
常林江
张会娟
叶文
吴才章
Chen Hongmei;Chang Linjiang;Zhang Huijuan;Ye Wen;Wu Caizhang(School of Electrical Engineering,Henan University of Technology,Zhengzhou 450001,China;National Institute of Metrology,China,Beijing 100029,China)
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2020年第7期136-145,共10页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(U1804161,61901431,51805148)
河南省科技攻关(172102210214,192102210066)项目资助
中国博士后科学基金(2020T130625)特别资助(站中)
关键词
协同导航
网络定位
环路和积数据关联
高斯滤波
延迟估计
collaborative navigation
network positioning
loop sum-product algorithm for data association
Gaussian filtering
delay estimation
