Robust Kalman Filters For Additive Gaussian-GSM Distributions

Authors: Gurui Jin, Pengcheng Hao, Zhenyu Liu, Xiao-Ping Zhang, Ercan E. Kuruoglu

This manuscript develops a unified robust Kalman-filtering framework for measurement noise modeled as an additive combination of Gaussian background uncertainty and a Gaussian scale mixture (GSM) component.

The work introduces a variance-offset formulation and develops MAP and LA-NM estimators for the associated latent-scale inference problem.