Information Gain-weighted Multi-sensor Arithmetic Average Fusion Kalman Filtering.

This paper presents a novel multi-sensor Kalman filter (KF) based on the arithmetic average (AA) fusion method. In this approach, the fusing weights are designed according to the online Kalman gain matrix obtained from each local filter. Both the standard KF and the unscented KF (UKF) are applied to linear and nonlinear state space models, respectively. Simulation results demonstrate the superior target tracking performance of our approach compared to the recently proposed suboptimal AA fusion method using the Kullback-Leibler divergence (KLD) in both linear and nonlinear scenarios.