基于EMD-小波随机消噪模型的GPS/INS组合导航

为提高GPS/INS组合导航精度,提出了EMD-小波INS随机误差削弱模型.首先分析了GPS/INS组合系统的误差特点,给出组合导航的基本原理.基于EMD的多尺度结构,结合小波分解理论,提出基于EMD-小波的GPS/INS组合导航技术路线.采集新南威尔士大学校园内的实测数据,分别选用1s与2s的GPS采样信息和100 Hz的INS数据,进行融合试验.用随机模型消噪前后的INS数据计算导航解,并与GPS测量的位置进行比较,分析误差分布情况.结果表明,滤波后的INS自主导航解要优于滤波前.当GPS更新周期为1 s时,X,Y,Z方向上导航精度分别提高9.07％,13.89％,35.14％,残差最大值分别减小2.62％,19.54％,31.23％.当GPS更新周期为2s时,X,Y,Z方向上导航精度分别提高17.82％,23.46％,22.48％,残差最大值分别减小14.34％,28.85％,17.48％.

EMD-wavelet based stochastic error reducing model for GPS/INS integrated navigation

The empirical mode decomposition(EMD)-wavelet model aiming to eliminate the inertial navigation system(INS) stochastic error is proposed to improve GPS(global positioning system)/INS integrated navigation accuracy.After analyzing the error characteristics of the integrated navigation system and introducing the basic principle of integrated navigation,a new GPS/INS integrated navigation route is plotted based on the EMD multi-scale structure and the wavelet decomposition theory.In order to carry out the fusion experiment,the GPS data which updates in 1 s and 2 s and the INS data in 100 Hz are gathered in the University of New South Wales campus.Navigation resolution computed by the INS data before denoising and after denoising is compared with GPS observation position.The error distribution of comparative information is analyzed further.The results show that the INS navigation resolution is better after filtering.When the update period of GPS data is 1 s,the navigation precision increases by 9.07%,13.89%,35.14% and the maximum of residual decreases by 2.62%,19.54%,31.23% in X,Y,Z direction,respectively.When the update period of GPS data is 2 s,the navigation precision increases by 17.82%,23.46%,22.48% and the maximum of residual decreases by 14.34%,28.85%,17.48% in X,Y,Z direction,respectively.