摘要
多雷达组网对空域运动目标跟踪时,组网雷达极坐标测量值与目标状态值呈非线性关系,不满足卡尔曼滤波线性化使用要求。提出将组网融合中心惯性坐标系虚拟为滤波观测坐标系,使得滤波状态和虚拟测量简化为线性关系,通过虚拟观测噪声建模、滤波初始化建模,解决了多雷达组网使用卡尔曼滤波对空域运动目标的最优化滤波估计问题。软件仿真测试和检飞数据验证表明:虚拟观测卡尔曼滤波算法(Virtual-Observation Kalman Filter Algorithms,VOKFA),滤波精度高、算法稳定性好,非常适合于多雷达组网滤波跟踪工程应用。
Under the circumstances of tracking targets in multi-radar networking(MRN), measure-value in polar coordinates of the networked radar(NR) has the nonlinear relation with state-value of targets tracking coordinates, which does not satisfy linear requirement of Kalman filter algorithm(KFA) application. Virtualizing the tracking coordinates of MRN as the measure coordinates of KFA was introduced. As a result, original nonlinear relation was simplified as a linear form. By means of modeling virtual observation noise and constructing the initialization strategy, KFA could solve state estimation problem in MRN. Simulation analysis and real data verification demonstrate that virtual-observation KFA(VOKFA) has quick computation velocity, high precision and good stability, which is very suitable for engineering application of data fusion in MRN.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2015年第4期851-858,共8页
Journal of System Simulation
基金
国家自然科学基金资助项目(61273001)
安徽省自然科学基金资助项目(11040606M130)
关键词
雷达组网
虚拟观测
坐标变换
误差统计特性
初始化策略
radar network
virtual observation
coordinate transformation
error statistical characteristic
initialization strategy
