摘要
对于测量噪声方差未知的捷联惯导系统(SINS),采用常规Kalman滤波进行初始对准会造成较大状态估计误差,甚至使滤波器发散。为了解决系统测量噪声方差未知或不确切知道时SINS的误差估计问题,提出一种基于随机逼近的自适应滤波方法。该方法将Robbins-Monro算法与Kalman滤波相结合,通过简化求逆运算,解决了系统观测噪声特性未知情况下SINS的误差估计问题,并提高了算法的数值稳定性。仿真结果表明,该方法能在系统测量噪声方差未知情况下有效实现SINS初始对准。
For the strapdown inertial navigation system（SINS）with unknown measurement noise covariance,applying conventional Kalman filter to initial alignment will lead to a large state estimation error or even filter divergence.To estimate SINS errors with unknown measurement noise covariance,an adaptive filter based on stochastic approximation is presented.In the filter,the Robbins-Monro scheme is applied to Kalman filter to solve the problem of SINS errors estimation with unknown measurement noise covariance,and the inverse operation is simplified to improve the numerical stability.The simulation results demonstrate the effectiveness of the adaptive filter in initial alignment for SINS.
出处
《传感技术学报》
CAS
CSCD
北大核心
2011年第7期1007-1010,共4页
Chinese Journal of Sensors and Actuators
基金
山东科技大学"春蕾"计划项目(2010AZZ049)
关键词
捷联惯导系统
初始对准
随机逼近
自适应滤波
测量噪声
strapdown inertial navigation system
initial alignment
stochastic approximation
adaptive filter
measurement noise
