改进的自适应衰减卡尔曼滤波算法

Improved Adaptive Fading Kalman Filter Algorithm

摘要:

SINS/GPS组合导航系统的融合算法主要是卡尔曼滤波,卡尔曼滤波实现最优估计的前提是系统的模型和随机噪声信息必须准确已知。实际情况下,大部分系统的模型和随机噪声信息不完全可知,这可能会导致滤波器估计精度下降。针对这一问题,根据求解遗传因子的方法不同对传统的自适应衰减卡尔曼滤波进行改进,提出一种改进的自适应衰减卡尔曼滤波。改进后的算法分别适用于系统噪声统计模型不准确可知和量测噪声统计模型不准确可知两种情况,分别对应于两种滤波算法,并且二者具有统一的滤波框架。仿真结果表明,改进的自适应衰减卡尔曼滤波比卡尔曼滤波精度较高,有效解决了因为噪声模型不准确导致的精度下降问题。

Abstract:

The fusion algorithm of SINS/GPS integrated navigation system is mainly based on kalman filter. Kalman filter is the optimal estimation on the conditions that system model and random noise information are accurately known. In practice, most system model and random noise information are not completely known, which may lead to filter estimation accuracy decline. Aiming at this problem, this paper improves the traditional adaptive fading kalman filter according to the method of solving forgetting factor, and proposes an improved adaptive fading kalman filter. The improved algorithm respectively applies in the cases that system noise statistical model cannot be accurately known and measurement noise statistical models cannot be accurately known, respectively corresponding to the two filter algorithms. What’s more, they have a unified filter framework. The simulation results show that the improved adaptive fading kalman filter is more accurate than kalman filter and it can effectively solves the accuracy decline problem caused by the inaccurate noise model.