Application of Kalman Filter method based on time series model in the landslide deformation forecast

Considering that the influence of environmental factors is different to deformation monitor points in different places of the landslide, some AR(n) models were preset, by means of calculation, AR(n) model whose residual standard deviation was small was chosen as the primary model, parameters of the primary model were looked as the status vector with dynamic noises to erect Kalman Filter Model. The example proved that the method can raise the fitting and forecast accuracy of the model.