Construct AKF to Improve Precision of Estimating and Predicting Dam Deformation Using Innovation

The Kalman filter(KF) method is propitious to process the dynamic dam deformation monitoring data.However,the state noise(R) and survey noise(Q) in the deformation monitoring is difficult to be provided precisely,the standard Kalman Filter method is confined.This article presents AKF based on innovation statistic properties to modify the R and Q,and to reflect correct statistic properties of current model by real time.With a dam survey data,this method succeed to overcome imprecision of stochastic model and deformation saltation to improve the precision of estimation and prediction in dam deformation.