Noise Covariance Estimation for Time-Varying and Nonlinear Systems

Kalman-based state estimators assume a priori knowledge of the covariance matrices of the process and observation noise. However, in most practical situations, noise statistics are often unknown and need to be estimated from measurement data. This paper presents a new auto-covariance least squares method for noise covariance estimation of linear time-varying and nonlinear systems. The new algorithm is more general than existing methods, since we do not assume asymptotic stability of the open-loop system trajectories.