一种基于冗余测量的自适应卡尔曼滤波算法

针对目前自适应滤波算法的不足,在测量系统量测噪声方差未知的情况下,设计了一种基于冗余测量的自适应卡尔曼滤波(RMAKF)算法。通过对系统冗余测量值的一阶、二阶差分序列进行有效的统计分析,可以准确估计系统量测噪声统计特性,进而在滤波过程中自适应调节噪声方差阵R,提高滤波精度。以全球定位系统/惯性导航系统(GPS/INS)松组合导航系统为对象进行了仿真实验,结果表明该算法在测量系统噪声特性未知或发生改变时,可对其进行准确估计,在采用低精度惯性器件情况下,滤波结果较其他主要自适应卡尔曼滤波算法有较明显的改进。

A redundant measurement adaptive Kalman filter algorithm

In order to solve the problems of current adaptive Kalman filter, this paper proposes a redundant measurement adaptive Kalman filter (RMAKF) in the situation that the measurement noise variance is unknown. This method could accurately estimate the statistical characteristics of measurement noise through calculating the first and second order difference sequences and adaptively tuning the variance matrix of measurement noise R in the process to improve the accuracy and precision. The simulation results show that when the algorithm is applied in GPS/INS loosely coupled integrated system, the proposed method is capable of estimating the noise variance when the statistical characteristic is unknown or changed. The simulation also shows that the filtering results has a great improvement compared with other adaptive Kalman filter when using low-accuracy inertial sensors.