基于方差分量估计的自适应融合导航  

Integrated Navigation by Using Variance Component Estimates of Multi-sensor Measurements and Adaptive Weights of Dynamic Model Information

作　　者：杨元喜[1] 高为广[2] 

机构地区：[1]西安测绘研究所,陕西西安710054 [2]信息工程大学测绘学院,河南郑州450052

出　　处：《测绘学报》2004年第1期22-26,共5页Acta Geodaetica et Cartographica Sinica

基　　金：国家杰出青年基金资助项目(49825107);国家自然科学基金资助项目(40174009和40274002)

摘　　要：多源传感器观测信息的精度一般不一致,利用方差分量估计理论可以合理地顾及各传感器观测信息在融合导航解中的权比。首先给出基于多源观测信息的融合导航解算模式,进而基于方差分量估计讨论了各传感器观测信息在融合导航解中的合理平衡问题,然后应用自适应因子顾及动力学模型预报信息的贡献,给出了计算过程,并利用模拟数据进行了试算与比较。The accuracy of the measurement vectors of the multi-sensor is usually different. In order to balance the contributions of the various navigation sensors to the last output, the principle of variance components estimation is applied. An integrated navigation estimator is given at first, based on the multi-sensor measurements. The variance components are applied for optimal balance of the contributions of the local navigation outputs that resulted from measurement vectors of the multi-sensors. On the basis of the integrated navigation output, the kinematic predicts are fused by using an adaptive filtering principle. An integrated navigation example by using simulated data is performed and analyzed.

关 键 词：方差分量估计 自适应融合导航 多传感器 动力学模型 

分 类 号：TP212.9[自动化与计算机技术—检测技术与自动化装置] P209[自动化与计算机技术—控制科学与工程]