基于中心差分卡尔曼滤波的初始对准方法研究  

Study on Initial Alignment Method Based on Central Difference Calman Filter

作　　者：贾鹤鸣[1] 杨泽文[1] 宋文龙[1] 

机构地区：[1]东北林业大学机电工程学院,哈尔滨150040

出　　处：《森林工程》2016年第6期66-70,76,共6页Forest Engineering

基　　金：中央高校基本科研业务费专项资金资助项目(2572014BB03);国家自然科学基金(31470714)

摘　　要：由于建立的大方位失准角下的捷联惯性导航系统误差模型具有非线性的特点,本文选用中心差分卡尔曼滤波(CDKF)方法对捷联惯导系统的控制对象进行初始对准仿真。并与无迹卡尔曼滤波(UKF)和扩展卡尔曼滤波(EKF)两种滤波方式进行比较,最后得出CDKF可以提高系统的图像和目标的滤波精度的结论,并且不需要具体模型计算出解析方程,节约了计算复杂的具有驱动约束方程的雅可比(Jacobian)矩阵的时间。最后的仿真结果证明CDKF对系统状态进行最优估计的算法明显优于EKF和UKF,同时具有较高的精度和收敛性,能够满足在非线性模型下的系统对导航的要求。Because the error model of the strapdown inertial navigation system is of nonlinear characteristics under the established large azimuth misalignment,the Central Difference Kalman Filter （CDKF） method was used to conduct the initial alignment simulation for controlling object of the strapdown inertial navigation system in this paper.Compared with the extended Kalman filter （EKF） and the unscented Kalman filter （UKF）,CDKF would improve the filtering precision of the system image and the objective,and did not need specific models to calculate the analytical equation,which saved the time for computing Jacobian matrix of driven complex constraint equations.The final simulation results demonstrated that the use of CDKF for the optimal estimation of system sate was better than EKF and UKF methods,the CDKF algorithm was of the high precision and convergence and could satisfy the system requirements for navigation under the nonlinear model.

关 键 词：大方位失准角 捷联惯导 初始对准 CDKF 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]