顾及系统噪声和观测噪声的分级自适应信息滤波算法

Hierarchical Adaptive Information Filtering Algorithm Considering System Noise and Observation Noise

高精度的载体动态导航与定位不仅需要对载体异常扰动和观测异常有良好控制,还需要对状态方程系统噪声及观测噪声的时变特性有准确认识和处理。首先针对包含系统噪声的动力学模型和包含时变观测噪声的导航系统,提出一种基于信息滤波形式的分级自适应滤波算法。然后针对系统噪声的渐变性和突变性,增加了遗忘因子和二段自适应因子,提高了对突变噪声估计的稳定性;顾及观测噪声的时变特性,采用传感器间差分和观测数据历元差分法估计观测噪声协方差。最后进行了仿真实验和深海拖体实验,结果表明,该算法不仅可以有效地估计系统噪声,还能准确地估计时变观测噪声的协方差阵,提高水下载体动态参数的估计精度。

High-precision carrier dynamic navigation and positioning requires not only good control of abnormal disturbance and observation abnormality,but also accurate recognition and processing of time-varying characteristics of noise and observation noise in equation of state system.Aiming at the system noise dynamics model and the time-varying observation noise navigation system,and based on information filtering,a hierarchical adaptive filtering algorithm is proposed.The unbiasedness and effectiveness of the new algorithm are proved.Considering the gradual and fast changes of system noise,the new algorithm adds forgetting factor or two-stage adaptive factor to improve the stability of noise estimation for catastrophic systems.In addition,considering the time-varying of the observation noise,two different difference observation data are used and the covariance of the observation noise is effectively estimated by using the highprecision equation of state.The simulation and experimental results show that the new filtering algorithm can not only estimate the system noise simply and effectively,but also estimate the covariance matrix of the observation noise effectively,which improves the accuracy of parameter estimation of dynamic system.