改进Allan方差的自适应滤波在姿态解算中的应用  

Application of Improved Allan Variance Adaptive Filtering in Attitude Solution

作　　者：刘春[1] 汪志宁 戴雷[1] 卫吉祥 刘滔 LIU Chun;WANG Zhining;DAI Lei;WEI Jixiang;LIU Tao(School of Electrical and Automation Engineering,Hefei University of Technology,Hefei 230009,China)

机构地区：[1]合肥工业大学电气与自动化工程学院,合肥230009

出　　处：《传感技术学报》2020年第5期682-687,共6页Chinese Journal of Sensors and Actuators

基　　金：合肥市北斗卫星导航重大应用示范项目(发改办高技(2014)2564)。

摘　　要：针对低成本的惯性导航系统精度不足从而导致姿态解算容易发散的问题,提出一种改进Allan方差的自适应卡尔曼滤波算法。在滤波之前,先用四元数改进型PID的互补滤波来融合数据,以抑制数据的波动,同时也加快了运算速率。在对噪声进行分析时,运用Allan方差的分析方法,并组合高斯牛顿优化算法,提高了姿态解算的精度,能够对姿态角实现短时间内的稳定以及精确的跟踪。实验结果表明,使用自适应Sage-Husa算法处理两种噪声时比标准卡尔曼滤波算法的精度提高了30%左右。使用改进的Allan方差自适应滤波比使用自适应Sage-Husa滤波算法精度提高了40%左右,该算法也可用于精确单点定位与伪距定位。Aiming at the problem that the low-cost inertial navigation system has insufficient precision,which can cause the attitude solution easy to diverge,an adaptive Kalman filter algorithm with improved Allan variance is proposed.Before filtering,the complementary filtering of the quaternary modified PID is used to fuse the data to suppress the fluctuation of the data,and also accelerate the operation rate.In the analysis of noise,the analysis method of Allan variance is used,and the Gauss Newton optimization algorithm is combined to improve the accuracy of attitude calculation.It can realize stable and accurate tracking of attitude angle in a short time.The experimental results show that when using the adaptive Sage-Husa algorithm to process two kinds of noise,the accuracy is improved by about 30%compared with standard Kalman filter algorithm.The improved Allan variance adaptive filtering isimprovedabout 40%better than the adaptive Sage-Husa filtering algorithm,The algorithm can also be used for accurate single point positioning and pseudorange positioning.

关 键 词：姿态解算 互补滤波 自适应卡尔曼滤波 PID控制 ALLAN方差 高斯牛顿算法 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]