Aircraft trajectory filtering method based on Gaussian‐sum and maximum correntropy square‐root cubature Kalman filter

Abstract

Abstract Aiming at meetiing the need to filtering flight trajectory data for aircraft testing, a novel adaptive cubature Kalman filter (CKF) is proposed based on the maximum correntropy and Gaussian‐sum in this paper. Firstly, based on the traditional CKF algorithm, we introduced a Gaussian‐sum method to approximate non‐Gaussian noise to get more accurate filtering results in view of the problem of reduced filtering accuracy caused by the inherent non‐Gaussian nature of the noise and the system non‐linearity. Secondly, the maximum correntropy criterion is introduced to solve further the problem of improving the filtering accuracy of the system in the case of non‐linearity. Simulation results and actual data verification showed that the Square‐root cubature Kalman filter algorithm based on the maximum correntropy and Gaussian‐sum has higher accuracy than traditional filtering algorithms, which verified the algorithm's effectiveness in the application.

Keywords

• adaptive systems
• machine learning