利用SVM的GPS/INS组合导航滤波发散抑制方法研究

为验证导航模型在GPS信号更新频率较低的情况下的导航能力,给出了GPS/INS组合导航滤波模型。载体运动过程中,分别使信号中断5s、10s、15s。通过实验得出,GPS信号中断时间过长(10s以上),GPS信号恢复后,Kalman滤波器会产生发散现象。引入支持向量机,提出利用SVM内插GPS信号提高信息更新频率消除组合导航滤波器的发散。结果表明,GPS信号中断时间过长导致组合导航系统滤波发散的情况下,通过SVM内插GPS数据提高GPS更新频率,可以有效地抑制滤波发散,提高导航的准确性。

A Method to Prevent GPS / INS Integrated Navigation Filtering Divergence Based on SVM

A GPS / INS integrated navigation model is introduced in order to verify the naviga-tion ability in the condition of low GPS signal update frequency. The GPS signal break period were chosen 5s, 10sand 15sduring the vehicle motion, which has demonstrated that long signal interruption interval (above 10s) will lead to the Kalman filtering divergence after GPS signal reconstruction. Considering that the integrated navigation model has favorable effect in high GPS signal update frequency, a new method is proposed to prevent the filtering divergence. The approach adopted extensively is called support vector machine (S VM) inter-polating the GPS signal based on vehicle movement discipline. These findings of the research have led the author to the conclusion that SVM is able to increase the GPS signal update fre-quency to prevent the filtering divergence caused by long GPS signal interruption interval and strengthen the navigation accuracy.