Adaptive compensation method for gravity disturbance based on time-varying markov model

1The gravity disturbance vector (GDV) is the source of significant errors in the positioning and orientation for high-precision inertial navigation systems. In order to enhance dynamic estimation accuracy of GDV and the autonomous navigation positioning accuracy of INS, a time-varying second-order damping differential Markov (TV-SDDM) model was established for characterizing high-frequency GDV variations along navigation trajectories. Subsequently, the estimation of TV-SDDM model parameters was achieved using the proposed parameter identification method based on least squares region interpolation. Finally, an adaptive Kalman filtering algorithm based on TV-SDDM model was established to dynamically estimate and compensate the GDV. Simulations and practical experiments are conducted to evaluate the performance of the proposed compensation method. The results of land-vehicle experiments show the improvements in northward and eastward positioning accuracies by 34.3% and 41.6%, respectively when compared with no gravity disturbance compensation. Compared with the two conventional methods, the horizontal positioning accuracies are improved by 13.1% and 25.6%, respectively.