Variance Behavior and Signification in Probabilistic Framework Applied to Vehicle Localization

The aim of this paper is to propose a discussion about the covariance signification and representation in probabilistic vehicle localization processes. Especially, we focus on its behavior along different trajectories (straight lines and curves). We discuss the interpretation, on one hand, of the mean (computed through a Monte Carlo simulation or estimated through an extended Kalman filtering process) and, on the other hand, of the uncertainty ellipses