卡尔曼滤波与H_∞滤波在GFSINS中的对比  

Comparison of Kalman filtering with H_∞ filtering in GFSINS

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,黑龙江哈尔滨150001

出　　处：《华中科技大学学报（自然科学版）》2010年第5期19-23,共5页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(60604019)

摘　　要：针对加速度计随机噪声影响无陀螺捷联惯导系统(GFSINS)角速度解算精度的问题,提出了一种滤波设计方案.基于十加速度计构型方式,利用加速度计输出构成的角加速度项作为状态方程,而对数法解算的角速度作为观测量,构建了卡尔曼和H∞滤波器.在改变加速度计随机噪声、状态初始估值和噪声颜色的情况下,仿真分析和对比了卡尔曼滤波和H∞滤波的效果.结果表明,H∞滤波不受加速度计随机噪声统计特性大小、噪声颜色的影响,比卡尔曼滤波有更强的鲁棒性和稳定性;H∞滤波后,角速度精度比对数法提高了2个数量级.The angular velocity calculating precision of Gyro-free strapdown inertial navigation system （GFSINS） was affected by random noise of accelerometer. Thus a filtering design scheme was put for- ward. Based on a ten-accelerometer configuration, the angular acceleration composed of accelerometer output was chosen as state equation, angular velocity got by lognormal algorithm was taken as measurement value. Then Kalman and H∞ filter were built. Simulation was conducted when the altering random noise of accelerometer, initial estimation value and noise color. Results show that H∞, filter isnrt affected by the characteristic and color of random noise, and has better robustness and stability than Kalman filter. Moreover, compared with lognormal algorithm, precision of angular velocity is improved by two orders of magnitude after H∞, filter.

关 键 词：无陀螺捷联惯导系统 角速度 卡尔曼滤波 H∞滤波 对数法 

分 类 号：V241.62[航空宇航科学与技术—飞行器设计]