权值动态分配的GFSINS多角速度加权平均融合算法  

Fusion algorithm of angular velocity weighted averages for GFSINS based on dynamically allocating weights

作　　者：郝燕玲[1] 刘志平[1] 

机构地区：[1]哈尔滨工程大学自动化学院,哈尔滨150001

出　　处：《中国惯性技术学报》2010年第1期16-21,共6页Journal of Chinese Inertial Technology

摘　　要：针对单一角速度算法解算精度不高,提出将最优加权平均法应用于GFSINS角速度解算中。针对实际应用中各种算法无法准确获取到角速度解算误差方差的情况,给出了一种对方差进行实时估计,再基于最优权值分配原则来动态分配权值的加权融合算法;对积分法、开方法、微分法和对数法四种角速度算法进行了有效融合,并完成了仿真。仿真结果表明,即使精度最低的积分法参与加权融合,也能进一步提高融合精度,同时还验证了加权平均法在系统角速度解算中的有效性,融合后,角速度精度比对数法提高了一倍。In view of low precision of single angular velocity algorithm, an optimal weighted average method is applied to the calculation of GFSINS angular velocity. Since the variance of angular velocity errors for each algorithm can＇t be accurately obtained in the practical application, a weighted fusion algorithm is given which estimates the variance in real time and allocating weights dynamically based on optimal weight allocation principle. Integral algorithm, extraction algorithm, differential algorithm and lognormal algorithm are effectively fused, and the simulation is made. The results show that fusion precision can be further improved even if with the integral algorithm. Moreover, the validity of optimal weighted average method is proved. After fusion, the angular velocity precision is one times higher than that by logarithmic method.

关 键 词：无陀螺捷联惯导系统 角速度 加权平均 权值动态分配 融合 

分 类 号：U666.1[交通运输工程—船舶及航道工程]