摘要
针对强机动性目标难以使用单一运动模型来描述,组网雷达(NR)的极坐标测量值与融合中心惯性(FCRNS)坐标的状态值呈非线性关系等问题,提出将FCRNS坐标虚拟为观测坐标(称为虚拟观测卡尔曼滤波),以满足采用卡尔曼滤波(KF)进行多雷达组网目标跟踪的线性化约束;使用协同转弯模型+匀速直线模型构建交互多模型,解决了多雷达组网空域机动目标自适应运动建模问题;通过虚拟观测误差协方差矩阵建模、初始估计建模,构建了交互多模型虚拟观测卡尔曼滤波算法(IMM-VOKFA),用于多雷达组网对空域机动目标的滤波跟踪.将提出的IMM-VOKFA用于多雷达组网的转弯机动目标跟踪中,并与交互多模型扩展卡尔曼滤波(IMM-EKF)进行了对比.仿真结果表明,IMM-VOKFA滤波精度高,机动自适应性强,计算稳定性高,工程可用性好.
Because a single model cannot describe maneuvering targets accurately and measured values in the polar coordinates of networked radar （NR） have a nonlinear relation with state values in the coordinates system of the fusion center of the radar networking station（ FCRNS）, we propose a strategy of transforming the tracking coordinates of FCRNS into virtual observation coordinates to satisfy linear constraints in the target tracking of multi-radar networkingusing （ Kalman filter） KF. We construct an interacting muhiple model that combines the coordinated turn model and the constant velocity model to adaptively model the motion of the airspace ma- neuvering targets in the multiradar networking. Then we propose an interacting multiple model-virtual observa- tion Kalman filter algorithm （IMM-VOKFA） to track airspace maneuvering targets by modeling the covariance matrix of virtual observation errors and the initial estimation. We use the proposed IMM-VOKFA to track the turning of a maneuvering target in a multi-radar networking system and compare it with interacting multiple model extended Kalman filter algorithm. Simulation results demonstrate that IMM-VOKFA has strong motor a- daptability, good calculating stability, and strong engineering effectiveness.
出处
《信息与控制》
CSCD
北大核心
2015年第1期15-20,共6页
Information and Control
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
虚拟观察噪声
交互多模型
卡尔曼滤波
radar networking
virtual observation noise
interacting mutiple model
Kalman filter
