简化平方根容积卡尔曼滤波的INS/GPS紧组合算法

INS/GPS Tightly Integrated Algorithm Based on Reduced Square-Root Cubature Kalman Filter

针对INS/GPS紧组合导航系统非线性模型解算的实时性问题,提出了一种用简化平方根容积卡尔曼滤波算法(RSCKF)提高紧组合导航系统运算速度的方法,它是在时间更新环节将平方根容积卡尔曼滤波(SCKF)简化,简化后直接用状态转移矩阵求取状态一步预测和预测协方差矩阵,避免了原算法中采用求容积点近似计算的复杂过程。仿真实验将RSCKF算法与SCKF滤波算法、扩展卡尔曼滤波算法(EKF)的结果对比。结果表明,RSCKF,SCKF两种算法的估计精度要明显高于EKF算法,而且在保证估计精度相当的情况下,RSCKF算法可大大降低系统运算量。

By focusing on the problem of the real time of nonlinear model calculation, the reduced square root cubature Kalman filter（RSCKF）/s proposed to use for improving the operating rate of IN S/GPS tightly integrated navigation system. This reduced arithmetic simplifies the square root cubature Kalman filter （SCKF） in time update step, and the state-transition matrix is directly used for calculating the one-step prediction matrix of state and covariance by avoiding the complex approximate calculation process of calculating cubature. During the simulation experiments, the RSCKF results are compared with SCKF and EKF results and show that the RSCKF and SCKF perform well than EKF. The RSCKF algorithm performs nearly the same as the SCKF by precision and can effectively reduce the amount of calculation.