一种有色噪声条件下的改进卡尔曼滤波算法

An improved Kalman filter algorithm under colored noise

以捷联惯导系统(SINS)/多普勒雷达组合导航系统为研究对象,针对多普勒雷达测速噪声为缺少先验信息的有色噪声的滤波问题,提出一种有色噪声条件下的改进卡尔曼滤波算法。首先,建立有色量测噪声条件下的SINS/多普勒雷达组合导航系统模型;然后,采用重构状态空间模型的方法将量测噪声为有色噪声的状态空间模型进行白化,消除有色噪声对导航定位滤波计算的不利影响;最后,对缺乏先验信息有色量测噪声下的滤波问题,设计了一种改进的卡尔曼滤波算法。仿真实验结果表明,与自适应卡尔曼滤波算法和有色噪声卡尔曼滤波算法相比,所提算法的平均水平定位精度分别提升了23.97%、21.01%。

Taking the SINS/Doppler radar integrated navigation system as the research object,aiming at the filtering problem of Doppler radar velocity noise under the condition of colored noise lacking prior information,an improved Kalman filter algorithm under the condition of colored noise is proposed.Firstly,the SINS/Doppler radar integrated navigation system model under the condition of colored measurement noise is established.Then,the method of reconstructing the state space model is used to whiten the state space model in which the measurement noise is colored noise,and the adverse effect of the colored noise on the navigation and positioning filtering calculation is eliminated.Finally,an improved Kalman filter algorithm is derived for the filtering problem under colored measurement noise without prior information.The simulation test results show that compared with the adaptive Kalman filtering algorithm and the colored noise Kalman filtering algorithm,the average horizontal positioning accuracy of the proposed algorithm is improved by 23.97%and 21.01%,respectively.