动态阿伦方差辅助的卡尔曼滤波算法在GPS_INS组合导航中的应用_韦官余.pdfVIP

Dynamic Allan Variance aided Kalman Filter in GPS/INS Integrated Navigation Wei Guanyu,Xu Bojian,Ding Yang Application and Development of Beijing Global Information Center,100094,5128# Haidian Beijing Postbox: wells0592029@163.com Abstract: For GPS/INS integrated navigation system, when using a low cost IMU(Inertial Measurement Unit), the classical Kalman filter usually can not get expected result, the main problem lies in the varying property of the process noise, which is critical to the performance of classical Kalman filter. In this paper, we firstly separate random noise from the output of gyroscopes and accelerometers original using a wavelet soft threshold de-noise method. Then, we track the properties of the separated noises in real time using Dynamic Allan variance. Finally, we update the process noise covariance matrix using the results of above analysis. Compared with classical Kalman filter in low cost GPS/INS integration, the simulation experiment showed a clear improvement in performance of the new algorithm. Keywords: Dynamic Allan variance; Wavelet de-noise; Kalman filter; Integrated Navigation GPS/INS wells0592029@163.com GPS/INSIMU Allan 1 [4][5] GPS/INS Lorenzo Galleani Patriza Tavella Allan[6]Allan [9] Allan