Process Noise Source Localization Using Kalman Filter

Due to complexity in the systems, spatial distribution of unmeasured process noise that is required for the controller and observer design are often unknown. In this study an innovations correlations approach developed in Kalman Filter theory is used to localize the process noise from output measurements. The approach calculates covariance matrices from analysis of resulting innovations from an arbitrary filter gain. Aim of this paper is to review the innovation correlations approach and to evaluate its performance for localization of the process noise. Numerical results suggest that the method can be effectively used for source localization of process noise as well as estimation of noise covariance matrices.