Randomly Weighted CKF for Multisensor Integrated Systems

Hua Zong,Wenhui Wei,Zhaohui Gao,Yongmin Zhong,Chengfan Gu

doi:10.1155/2019/1216838

Abstract

The cubature Kalman filter (CKF) is an estimation method for nonlinear Gaussian systems. However, its filtering solution is affected by system error, leading to biased or diverged system state estimation. This paper proposes a randomly weighted CKF (RWCKF) to handle the CKF limitation. This method incorporates random weights in CKF to restrain system error’s influence on system state estimation by dynamic modification of cubature point weights. Randomly weighted theories are established to estimate predicted system state and system measurement as well as their covariances. Simulation and experimental results as well as comparison analyses demonstrate the presented RWCKF conquers the CKF problem, leading to enhanced accuracy for system state estimation.

Highlights

Nonlinear system state estimation is an importance research topic in various fields such as multisensor data fusion and integrated navigation, weapon guidance and control, intelligent transportation, system identification, and target tracking
Latitude errors 4.9886 8.3168 2.3516 3.6235. It is evident from the above simulation and experimental results and analyses that randomly weighted CKF (RWCKF) is capable of restraining system error’s disturbance on system state estimate, resulting in enhanced navigation accuracy comparing to cubature Kalman filter (CKF)
A RWCKF is presented to address the inherent problem of CKF by incorporating random weights in CKF

Summary

Introduction

Nonlinear system state estimation is an importance research topic in various fields such as multisensor data fusion and integrated navigation, weapon guidance and control, intelligent transportation, system identification, and target tracking. Since the adaptive factor is determined empirically, this method cannot adapt to the uncertain nature of system error, resulting in suboptimal or biased solutions It still suffers from the problem due to the use of arithmetic mean for state and measurement predictions. Different from our previous work [32], this method directly modifies cubature point weights based on prediction errors, rather than estimates system noise statistics, to restrain system error’s influence on system state estimation. If there exists a system model error, the CKF estimation will be deteriorated It is evident from the above that the predicted state and measurement as well as their covariances are computed by arithmetic mean via the same weight 1/m.

Randomly Weighted CKF

Performance Evaluation and Discussion

Conclusions

Conflicts of Interest