Rao-Blackwellized Variational Bayesian Smoother for Mobile Robot Localization

This paper proposes a Rao-Blackwellized variational Bayesian smoother for mobile robot localization with unknown measurement noise covariance. The measurement noise covari-ance is considered as a random matrix to be estimated where inverse Wishart distribution is used to describe the uncertainty of the measurement noise covariance. Variational Bayesian approxi-mation is applied to calculate the distribution of the measurement noise covariance. Based on landmark measurements obtained by the mobile robot, the robot pose, landmark locations, and measurement noise covariance are estimated by the proposed method in which the estimated robot pose and landmark locations are obtained by a Rao- Blackwellized estimator. The numerical simulation is provided to show the effectiveness and advantage of the proposed method.