Robust Kalman Filter For Additive Gaussian-Student’s t Distribution
Authors: Gurui Jin, Pengcheng Hao, Zhenyu Liu, Xiao-Ping Zhang, Ercan E. Kuruoglu
This work studies robust state estimation when measurement noise contains both a Gaussian component and a heavy-tailed Student’s t component. The proposed RKF-AGST explicitly models this additive hybrid-noise structure within a Kalman-filtering framework.
In the target-tracking simulations summarized in my current CV, the method achieved a 12.27% RMSE reduction relative to the standard Kalman filter.