Robust Cubature Kalman Filter for Power System State Estimation With Multi‐Rate Measurements

ABSTRACT

In this paper, a multi‐rate robust cubature Kalman filter (MR‐RCKF) algorithm is proposed to address the multi‐rate issue caused by the varying sampling times of multiple sensors and the adverse impact of the changing noise distribution on the perception accuracy. First, a multi‐rate mechanism based on the assumption of non‐integer multiples is proposed, and the enhanced covariance intersection fusion method is utilised to handle the multi‐source information fusion problem. Then, the generalised correntropy loss (GCL) criterion is introduced into the cubature Kalman filter (CKF) to handle the changed non‐Gaussian noise. Moreover, an adaptive kernel bandwidth strategy based on measurement covariance and innovation is designed to optimise the change rate of the estimated values. Finally, the comparison results with other Kalman‐type estimators show that the MR‐RCKF has more robust estimation performance under the influence of non‐Gaussian noise.