Extended target probability hypothesis density filter based on cubature Kalman filter

Aiming at the extended target tracking problem in a non‐linear Gaussian system, we proposed an extended target probability hypothesis density (EPHD) filter based on the cubature Kalman filter (CKF). To approximate the analytical solution of the extended target tracking, a spherical radial cubature rule was applied to make it possible to numerically compute multivariate moment integrals in the non‐linear Bayesian filter. Cubature points and weights were obtained to approximate the integrals in the process. The new algorithm achieved almost the same filtering accuracy as the Gaussian mixture extended Kalman EPHD (EK‐EPHD) filter, when solving tracking problems in such complex conditions that the Jacobian matrix of a non‐linear function does not exist or is difficult to solve. This work provides a new approach for the extended target tracking under the non‐linear Gaussian system.