自适应交互集合Kalman滤波的动态精密单点定位  

Dynamic Precise Point Positioning Based on Adaptive Interacting Ensemble Kalman Filtering

作　　者：王双红[1] 聂建亮[2] 

机构地区：[1]中原工学院电子信息学院,河南郑州450007 [2]国家测绘局大地测量数据处理中心,陕西西安710054

出　　处：《自动化仪表》2013年第5期10-12,共3页Process Automation Instrumentation

基　　金：国家自然科学基金资助项目(编号:61074022)

摘　　要：为了进一步提高动态精密单点定位的解算精度,采用交互多模型思想构建动态精密单点定位解算模型。采用集合Kalman滤波算法降低GPS数据中因非高斯噪声正态化处理造成的精度损失;利用状态预测向量残差信息,采用自回归模型(AR)修正当前历元的预报值,提高动力学模型的可靠性;根据单位权中误差自适应选取最终滤波解。对某载GPS数据进行验证,计算结果表明,自适应交互集合Kalman滤波是一种性能可靠、精度高的滤波算法。In order to further improve the solving accuracy of dynamic precise point positioning （ DPPP}, by adopting the concept of interacting multiple model, the solving model of DPPP is set up. By using ensemble Kalman filtering algorithm, the accuracy loss caused by normality processing of the non Gaussian noise in GPS data is reduced. With help of predicated state vector residuals information, and applying autoregressive （AR） model, the current epoch forecast value is corrected to enhance the reliability of dynamics model, the final filtering result is selected in accordance with error adaptation of unit weight. Using certain onboard GPS data for validation, the result of calculation indicates that the adaptive interacting ensemble Kalman filtering is a reliable and high precise filtering algorithm.

关 键 词：交互多模型 卡尔曼滤波 GPS 非线性 数据采集 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]