GPS/INS组合导航中两步自适应滤波方法

Two-Step Adaptive Filtering Method on GPS/INS Integrated Navigation

GPS/INS组合导航实际应用中存在系统模型偏差、噪声模型不确定等问题,导致卡尔曼滤波器无法实现最优滤波效果,严重时甚至导致滤波发散。渐消卡尔曼滤波器和自适应卡尔曼滤波器通过引入单渐消因子和单自适应因子可以部分解决上述问题,但是不足在于单因子只能进行整体调整,不能精确调整各个通道。针对此问题,本论文提出一种2步自适应卡尔曼滤波算法,构造基于残差协方差估计的多重渐消因子和自适应因子对各个通道精确调整,克服动态环境下跟踪性差的局限性。实验结果表明,改进后的自适应卡尔曼滤波算法可以精确调整各通道,增强系统的定位精度、跟踪性能和鲁棒性。

There are declination and uncertainty with measurement noise in the GPS/INS integrated navigation system in practical,which result in difficulty to obtain optimal filtering result even by diverging with Kalman filter. The fading Kalman filter（ FKF） and adaptive Kalman filter（ AKF） can partially solve the above problems by introducing the scalar fading factor and adaptive factor,but they can only adjust the whole result rather than the each channel exactly. In this paper,a two-step adaptive Kalman filter proposed by multiple fading and adaptive factors based on innovation covariance estimation is established to adjust each channel exactly for reducing the influence of ambient noise on the system and overcoming the limitations of tracking poorly in a dynamic environment. The experiment and simulation results show that the improved adaptive Kalman filtering algorithm can adjust each channel exactly and improve the system accuracy,tracking performance and robustness.