A new maneuvering target tracking method using adaptive cubature Kalman filter
Yiou Sun, Jingwen Xie, Junhai Guo · 2014
The cubature Kalman filter algorithm needs to know the statistical characteristic. When tracking a target, the algorithm may result in a divergence because of an unknown noise. This paper proposes an adaptive cubature Kalman filter based on cubature Kalman filter and Sage-Husa estimator. The proposed algorithm brings a Sage-Husa estimator in the cubature Kalman filter algorithm, so it can estimates the statistical parameters of unknown system and observation noise in real time, refrain the algorithm from divergence. The proposed algorithm can also decrease the tracking error due to unknown noise, and increase the accuracy and numerical stability effectively. According to the simulation result, the ACKF algorithm has a satisfactory performance, and has better accuracy and numerical stability comparing with UKF algorithm and CKF algorithm.