Modified Adaptive Federated Student's t Maximum Correntropy Criterion Variational Adaptive Kalman Filtering for Multi-Source Data Fusion

Abstract

The traditional federated Kalman filter-based multi-source data fusion algorithm performs poorly in the presence of outliers and unknown noise, a modified federated robust Student's t maximum correntropy criterion variational adaptive Kalman filter is proposed in this paper to tackle the issue. First, an improved robust Student's t maximum correntropy criterion variational adaptive Kalman filter is proposed for local estimation. The algorithm introduces an adaptive factor in the Student's t variational adaptive Kalman filter algorithm to correct the bias of the error covariance matrix, which improves the estimation accuracy of the algorithm. In addition, an improved kernel width based on the maximum correntropy criterion is employed for modifying the correntropy gain to adjust the filtering gain of the algorithm. Second, an improved adaptive information-sharing factor is developed to adaptively regulate the fusion weight of the local sensor based on the estimation accuracy of the local filter. Finally, the simulation verifies that the proposed algorithm has higher estimation accuracy than other existing algorithms.