Adaptive weighted federated Kalman filtering based on Mahalanobis distance and its application in navigation
Yi Gao,Zhaohui Gao,Hua Zong,Shesheng Gao,Genyuan Hong
DOI: https://doi.org/10.1049/cmu2.12664
IF: 1.345
2023-08-28
IET Communications
Abstract:This paper proposes the adaptive weighted federated Kalman filtering based on Mahalanobis distance (MD‐AWFKF). Firstly, by calculating the Mahalanobis distance between the predicted value and the measured value of the system state, the random fluctuation of the measured value is detected. The statistical characteristics of the system measurement noise are adjusted at any time according to random fluctuations in the measured value. Secondly, it constructs an adaptive amplification factor to dynamically adjust and suppress the measurement noise in the subsystem, and reduce the impact of measurement information contamination in subfilters on the main filter. And then, by introducing an adaptive federated information distribution coefficient, using the federal Kalman filtering method for global information fusion to reduce the influence of inaccurate estimation of subfilters on the main filter. Finally, the simulation proves that the filtering performance of the proposed MD‐AWFKF is better than the traditional Kalman filtering algorithm, and the accuracy of navigation filtering calculation is improved. Due to the federal Kalman filter is used to directly fuse the measurement information into the main filter without processing, resulting in the problem of reduced filtering accuracy. An adaptive weighted federated Kalman filtering based on Mahalanobis distance was proposed in this paper. By calculating the Mahalanobis distance between the predicted value and the measurements of the system, the random fluctuation of the measurements is detected. The statistical characteristics of the system measurement noise are adjusted at any time according to random fluctuations in the measurements. And then by using a adaptive amplification factor to dynamically adjust the measurement noise in the subsystems, and reduce the impact of measurement information contamination in subfilters on the main filter. The adaptive federated information distribution coefficient is used to realize the global information fusion of the federal Kalman filter method, to reduce the influence of inaccurate estimation of subfilters on the main filter.Simulation results and comparison analysis prove that the filtering performance of the proposed is better than the traditional federated Kalman filter (FKF) and adaptive FKF, which can improve the accuracy of the integrated navigation system.
engineering, electrical & electronic