变增益联邦KF组合导航定位算法研究

Research on variable gain federation Kalman filter location algorithm of coupled system

GPS接收模块解算出的伪距误差是GPS/INS组合导航系统的主要误差,采用一种二级联邦卡尔曼滤波组合导航算法加以削弱,将卫星接收模块解算出的伪距信息和多普勒频移信息在第一级卡尔曼滤波后,再通过主滤波器与INS模块解算出的信息进行修正处理,得到校正量和定位位置最优估计。随着滤波步数增加,系统预测误差方差阵逐渐趋于零,状态估计会过分依赖旧量测值,从而导致滤波发散,影响系统定位精度。为有效提高新量测值的修正作用,在联邦卡尔曼滤波组合导航算法中引入一种可变加权系数。仿真结果表明,改进后的变增益联邦卡尔曼滤波算法具备联邦卡尔曼滤波的优点,并且该算法滤波效果有较明显的改善,能有效抑制滤波发散,提高系统的定位精度。

The pseudorange error worked out by the GPS receiver module is the major errors of GPS/INS navigation system, which can be weakened by using one federal Kalman filter algorithm of integrated navigation. This algorithm has two Kalman filters. The first Kalman filter filters the pseudo range and Doppler shift got from the GPS module. Then, the results got from the first Kalman filter is filtered by the second Kalman filter with the data got from the INS module. And it gets the correction and the optimum estimate. With the filter step increasing, the system prediction error variance tends to zero gradually, then the state estimation will excessively dependence on the old measurements, which will cause filter divergence and affect the position accuracy. To effectively increase the correction amount of the new measurements, this paper introduces a variable weighting factor in federal Kalman filter integrated navigation algorithm. The simulation results show that the improved variable gain federal Kalman filtering algorithm has the advantages of federal Kalman filter, and the filtering effect of this algorithm is improved obviously, which can effectively restrain the filter divergence and improve the positioning precision of the system.