Colored noise estimation algorithm based on autocovariance least-squares method

The statistic character of colored noise directly affects the application of the Kalman filter in the actual dynamic system. In this paper we propose a colored noise estimation algorithm based on autocovariance least-squares method. The Kalman filter algorithm with the colored noise driven by non-Gaussian white noise is given. Then the autocovariance linear equations are derived under the conditions that the process noise and the measurement noise are all colored noises, and the least-squares method is used to solve the autocovariance linear equations and the covariances of the colored noise are estimated. The simulation results show the correctness and validity of the proposed algorithm.