Research On Data Processing Methods Of SINS/GPS Integrated Navigation

This dissertation mainly focuses on the research of data processing method of SINS/GPSintegrated navigation, which includes SINS navigation algorithm, data processing model ofSINS/GPS integrated navigation, kalman filter, robust Filter algorithm and fault detectionalgorithm of SINS/GPS integrated navigation, smoothing filter algorithm and so on. The mainworks and new standpoints are listed as follows:1. The navigation algorithm of strapdown inertial navigation system in the local navigationframe is researched, and the calculation of the rotation vector navigation algorithm issummarized, as well as the main error sources of the system.2. The error equations of strapdown inertial navigation system in the local navigation frameare derived, and the position error equations are improved. The mathematic model of looselycoupled integration and tight coupled integration are studied, and the specific form of stateequations and measurement equations are given, besides the loosely coupled integration andtightly coupled integration based on extended kalman Filter algorithm are performed with actualexperiment data. The results show that: the precision of loosely coupled integration and tightlycoupled integration are equal when the visible satellites are more than4, but the tightly coupledintegration performs better when the visible satellites are less than4.3. The robust kalman filter algorithm is researched aimed at the abnormal GPS observations.Firstly, the effects of abnormal observations in the integrated navigation system are analyzed,and then the robust kalman filter based on the robust least square estimation algorithm isresearched, in which the equal weight factors and equal weight covariance are constructed by theleast square residual statistic. The robust kalman filter is used in the data processing of tightlycoupled SINS/GPS integration. The calculation results show that: robust kalman filter algorithmcan reduce the effect of the abnormal observations to the navigation precision when the visiblesatellites are more than four, but it degrades to normal kalman filter while the visible satellitesare less than four.4An extended robust kalman filter based on innovation chi-square test algorithm ispresented aimed at the problem of no redundant observations and abnormal observations. Theinnovation chi-square test algorithm is used to detect the innovation vector, and the abnormalobservations are classified to the random model. The equal weight factors and equal weightcovariance are constructed according to the difference between the statistics and the designthreshold. The extended robust kalman filter based on innovation chi-square test algorithm isused for the performance of loosely coupled integration and tightly coupled integration withactual experiment data. The calculation results show that: the extended robust kalman filterbased on innovation chi-square test algorithm can also effectively weaken the influence of theabnormal observations to the precision of the navigation enven with no redundant observations,and improve the precision of the integrated navigation system, so that the integrated navigationsystem can be used in harsh navigation environment.5. In the data post-processing of SINS/GPS integrated navigation, the high precisionsmoothing algorithm is researched aimed at the problem of precision degradation caused by GPSreceiver loss of lock for a long time. The significance in the theory of optimal smoothing isanalyzed during the GPS signal interrupted, and the RTS smoothing algorithm equations aregiven, at the same time three sets of experimental program are designed, and the EKF algorithmand RTS algorithm are used for the verification and analysis with actual experiment data during the case of GPS without the loss of lock and continuous loss of lock for300seconds and600seconds. The results show that: RTS is not only able to play a smoothing effect to the navigationsolution results, but also can significantly weaken the influence of GPS loss of lock to theintegrated navigation system.