A novel CKF method for target tracking

Abstract

This paper presents a new target tracking method. The presented method which named marginalized cubature Kalman filter is based on standard cubature Kalman filter and marginalized moment estimator. The marginalized moment estimator uses sigma-points sampling and Guass-Hermite integration to estimate the mean and covariance. The proposed algorithm which is called MCKF in short, uses marginalized moment estimator to calculate the state's mean and covariance in the CKF framework and gets a better accuracy and keep the covariance matrix being positive definite. Simulation indicates the presented algorithm's feasibility and improved performance.