Decomposition method of covariance matrix for command and control system

Abstract

The invention discloses a covariance matrix decomposition method for a command and control system. The method comprises the following steps: S10, acquiring acquired target track information; s20, fusing the target track information to obtain a covariance matrix matrixP of the target track information; s30, after the covariance matrix matrixP is preprocessed, whether an abnormal line exists in the covariance matrix matrixP or not is judged; and S40, the covariance matrix matrixP with the abnormal row is corrected, and matrix decomposition is carried out on the corrected covariance matrix matrixP. According to the decomposition method, decomposition can still be carried out under the non-positive definite condition of the matrix, the requirement for calculating the Sigma point set by the UKF filter is met, and the purpose of improving the robustness of the track filter of the command and control system is achieved.

Description

technical field [0001] The invention relates to the technical field of data processing. More specifically, it relates to a method for decomposing the covariance matrix of a command and control system. Background technique [0002] In the command and control system, it is necessary to accurately correlate the track information generated by observations of multiple information sources, and fuse them to generate a comprehensive track to generate a unified air intelligence situation. Among them, the Kalman filter method is a commonly used fusion method. KF (Kalman filter), EKF (extended Kalman filter), UKF (unscented Kalman filter) and other linear and nonlinear filtering methods, as a powerful tool for real-time state estimation, have been widely used in various fields of engineering practice. [0003] The core idea of ​​UKF is to approach the nonlinear function distribution through sampling strategy, use the calculated Sigma point set to approximate the real probability densi...

AI Technical Summary

Problems solved by technology

[0004] The premise of the Cholesky decomposition of the covariance matrix is ​​that the matrix is ​​positive definite, but in the actual measurement process of the information source, due to the constraints of the measurement environment, target state, equipment capability and other conditions, the covariance matrix may not be positive definite in the filtering process. If the situation causes the filter to crash, the common way to deal with it is to restart the filter, but this will lose the historical observation information

Images