A Novel Data Fusion Method Based on Signal's Reliability
Ming Xia Xiao, Liang Pan, Yan Bu, Li Li Wan · Applied Mechanics and Materials · 2013
The traditional Kalman filter is able to obtain the optimal estimation of the estimated signals. However, it fails to consider their reliability. In real applications, the estimated signals may include outliers. Fortunately, we are able to know the reliability of the signals transcendentally. In this paper, we derive the one-dimensional data fusion formulas based on signals reliability which is according to minimum variance restriction. Furthermore, a corresponding data fusion scheme is proposed. Experimental results show the propose data fusion method performs much better than traditional methods.