Updating Noise Parameters of Kalman Filter Using Bayesian Approach

摘要: Kalman filter [1] is known to be a robust tool in state estimation for linear or slightly nonlinear systems diverse disciplines of science and engineering. In the area structural engineering, has also received enormous attention over years due its importance model updating, response prediction, control health monitoring. The algorithm becomes popular since it provides not only but associated uncertainty estimation. addition, online so that vector immediately updated once new data point obtained. However, accuracy depends on prior knowledge process noise measurement parameters, which difficult obtained practice. present study, Bayesian propabilistic approach [2] proposed estimate these parameters case when input zero-mean Gaussian white limited output measurements are available. optimal estimates chosen by maximum likelihood criterion. Through two illustrative examples, estimated close actual values sense located region with significant probability density. Therefore, concluded able provide accurate hence estimation, filter.