关联概率密度加权重力异常UKF滤波匹配导航算法  

The UKF Matching Algorithm Using Gravity Anomaly Based on Correlative Probability Density Add-Weight

作　　者：王伟[1] 李姗姗[1] 邢志斌[1] 曲政豪 郑伟 

机构地区：[1]信息工程大学,河南郑州450001 [2]工程兵学院,江苏徐州221004

出　　处：《测绘科学技术学报》2015年第4期349-352,356,共5页Journal of Geomatics Science and Technology

基　　金：国家自然科学基金项目(41274029);国家863计划项目(2013AAl22502);信息工程大学地理空间信息学院硕士学位论文创新与创优基金资助(S201407)

摘　　要：无味卡尔曼滤波(UKF)是一种通过采样来近似非线性分布,从而对非线性问题进行次优估计的滤波方法。利用实时观测得到的重力异常以及航行区域参考重力异常图,可以建立基于重力异常的UKF滤波匹配导航算法,以此校正惯性导航系统的漂移误差。针对选取与UT变换相同权系数来求取预测观测值而可能导致求得的预测观测值为伪观测值的问题,提出了利用关联概率密度函数对采样观测值进行加权的重力异常UKF滤波匹配导航算法。通过对某实验区域的实验进行计算分析,结果表明,基于关联概率密度加权重力异常UKF滤波算法能够克服传统加权预测观测值带来伪观测信息的问题,将惯性导航系统经纬向漂移误差降低至1.1 n mile以内,均优干传统加权算法和纯惯性导航系统的定位精度。Unscented Kalman filtering is used to estimate the nonlinear problems through sampling approximating a nonlinear distribution. The real time observed gravity anomaly data and gravity anomaly referenced map in the navi- gation area can be used to correct the drifting errors of inertial navigation system based on the UKF matching navi- gation algorithm. For the problem that the false observations might be arose by choosing the same weight coefficient as the UT transformation, the gravity anomaly UKF filtering algorithm from the view of probability density to add weight for the sampling data has been researched. The experiment has been made in some area, the results showed that the problem mentioned above could be overcome based on the correlative probability density add-weight, the drifting errors of inertial navigation system in longitude and latitude could be reduced within 1.1 n mile, the posi- tioning accuracy of the navigation system was better than the traditional algorithm and the INS.

关 键 词：关联概率密度 重力异常 无味卡尔曼滤波 加权 非线性 

分 类 号：P223[天文地球—大地测量学与测量工程]