Abstract:In the kinematic positioning or dynamic deformation monitoring, random walk model, constant velocity model and constant acceleration model are usually applied as the state model of Kalman filter. However, with these simple kinematic models to describe the complex movement of the deformed body,the modeled results will are inevitably contained errors and under some abnormal disturbance there are inevitably significant model biases in the original kinematic model, which will seriously affect the accuracy of filtering. With simulated data random walk model, constant velocity model, constant acceleration model and constant velocity model with correction of the model biases are comparatively analyzed. The advantages and disadvantages of the above models are summarized for the application of deformation monitoring.