Composite embedded cubature Kalman filter

Abstract

Summary The embedded cubature Kalman filter (ECKF) is proven as a kind of algorithm that has higher precision than cubature Kalman filter. Based on the ECKF, a new algorithm, named composite ECKF (CECKF), is presented in this paper. The new CECKF can increase filter precision by means of reusing the embedded cubature rule in the process of numerical approximation, which can realize higher precision than ECKF in theory. At the same time, another new algorithm, named adaptive CECKF, is also put forward by introducing an adaptive method based on CECKF. Through the comparison of simulation, the proposed methods are able to enhance filter precision.