多传感器加权观测融合自适应UKF滤波器  

Adaptive Weighted Measurement Fusion Unscented Kalman Filter for Multisensor System

作　　者：郝钢[1,2] 叶秀芬 

机构地区：[1]哈尔滨工程大学,哈尔滨150001 [2]黑龙江大学,哈尔滨150080

出　　处：《宇航学报》2011年第6期1400-1408,共9页Journal of Astronautics

基　　金：教育部科学技术研究重点项目(209038);黑龙江省自然科学基金(F201015)

摘　　要：对于带有相同观测方程和未知噪声统计的非线性多传感器系统,提出了一种基于Sage-Husa估计的自适应UKF滤波算法。该算法利用导出的平稳随机序列的相关函数估计系统观测噪声方差统计R(j),并证明了其收敛性。进而利用Sage-Husa估计算法得到自适应UKF滤波算法。该方法避免了传统Sage和Husa的自适应滤波算法不能处理Q和R均未知的系统的局限性。为了将多传感器信息加以充分利用,提高滤波精度,本文利用加权最小二乘法(WLS),实现了多传感器加权观测融合自适应UKF滤波器。一个带3传感器非线性系统的仿真例子说明了该算法的有效性。For the multisensor nonlinear systems which have the same measurement function,an adaptive unscented Kalman filter is presented based on the Sage-Husa estimator.This algorithm can estimate the measurement noise variances R（j） of the subsystems by the correlated functions matrix of these educed sequences,and its convergence is also proved.The algorithm avoids the disadvantage of classic Sage-Husa estimator when the Q and R are all unknown.To take full advantage of the information of multisensor systems and improve the filtering accuracy,the adaptive weighted measurement fusion unscented Kalman filter is obtained by using the weighted least squares（WLS） method.A simulation example for a nonlinear system with 3 sensors shows its effectiveness.

关 键 词：自适应滤波 Unscented卡尔曼滤波器 加权观测融合 

分 类 号：O211.64[理学—概率论与数理统计]