Adaptive cubature Kalman filter with directional uncertainties [Correspondence

Abstract

In this paper, an adaptive cubature Kalman filter (ACKF) is proposed to improve the performance of the conventional cubature Kalman filter. The ACKF uses a new cubature rule that combines a third-degree spherical rule with an adaptive higher degree radial rule along the directions of larger uncertainty. More accurate and robust results can be obtained with slightly more cubature points than the conventional third-degree cubature Kalman filter (CKF). Compared with other high-degree Gaussian filters, ACKF uses much fewer points but maintains very close performance. A target tracking problem is used to demonstrate the effectiveness of the proposed filter.