基于多元插值的DD型滤波算法仿真研究  

Simulation of DD Type Filter Algorithm Based on Multi-dimensional Interpolation

作　　者：李景熹[1] 王树宗[1] 王宇 

机构地区：[1]海军工程大学海军兵器新技术应用研究所,湖北武汉430033 [2]总参陆航部军代局财务审价处,北京100083

出　　处：《计算机仿真》2006年第6期124-126,共3页Computer Simulation

摘　　要：DD型滤波是一种基于Stirling多元插值方法,将函数按多项式近似展开的非线性滤波算法。相对于扩展卡尔曼滤波而言,它不需要对非线性函数进行微分运算,具有滤波精度高、数值稳定性好和适用范围广的优点,其运算量却与扩展卡尔曼滤波相当。对DD型滤波算法进行了深入分析,并将该算法应用于状态估计领域。对一多传感器目标跟踪问题进行了仿真计算,仿真结果表明了DD型滤波算法的有效性和实用性。The filter of DD type is a kind of nonlinear filter algorithm based on polynomial approximations according to multi - dimensional Stifling interpolation method. In contrast to extended Kalman filter, the filter needn＇t obtain nonlinear function＇s differential coefficient, and is characterized by high precision, good reliability and extensive applicability. Moreover, its computational burden is equivalent to extended Kalman filter . The paper analyses this kind of filter and applies it to the state estimation field. A problem of targets tracking by multi - sensor is presented, simulation results show filter algorithm＇s validity and practicability.

关 键 词：非线性滤波 扩展卡尔曼滤波 差分算法 插值 状态估计 

分 类 号：O211.64[理学—概率论与数理统计]