基于自适应滤波的惯性+多模卫星组合导航算法研究  

Research on an INS/ Satellite Integrated Navigation Algorithm Based on Adaptive Filtering

作　　者：夏巍巍[1] 牟建华[1] 傅志民[1] 暴飞虎[1] 

机构地区：[1]第二炮兵装备研究院,北京100085

出　　处：《导弹与航天运载技术》2014年第6期12-16,共5页Missiles and Space Vehicles

摘　　要：惯性+多模卫星组合导航系统通常采用定常参数的加权最小二乘算法进行多传感器信息融合,若加权系数与实际噪声统计特性不吻合,将会对组合导航精度产生不利影响。为解决该问题,提出一种基于自适应滤波的改进信息融合算法,对惯性及卫星导航数据应用自适应Kalman滤波以估计惯性导航误差,对滤波输出进行基于加权最小二乘法的多模信息融合,并根据滤波误差方差阵的解算结果对加权系数进行实时调整以优化估计精度。仿真结果表明该算法能够在一定程度上提高组合导航系统的精度和对不同随机噪声的适应能力。Weighted Least Square （ WLS ） is a primary algorithm of INS/Satellite integrated navigation system for information fusion. The navigation accuracy will be reduced if the weighted coefficient is not consistent with the real statistic characteristic of the measuring noise. An information fusion algorithm based on adapted filtering is presented to solve this problem. Adaptive Kalman Filtering is used to estimate the INS error, and estimate~ from multimode systems are fused using WLS, while the weighted coefficient is adjusted according to the filtering error variances estimation. It is validated by simulation that the algorithm can improve the precision of integrated navigation system and the adaptability to different stochastic.

关 键 词：组合导航 加权最小二乘 自适应滤波 信息融合 

分 类 号：V448.22[航空宇航科学与技术—飞行器设计]