Onto computing the uncertainty for the odometry pose estimate of a mobile robot

Solving the navigation issue for a mobile robot in a 2D space requires using internal and external sensors, so researchers try to fuse data from different sensors using methods as for example Kalman filtering. Those methods need an estimation of the uncertainty in the pose estimates obtained from the sensory system, usually expressed by a covariance matrix and obtained from experimental data. In a previous work, a general method to obtain the uncertainty in the odometry pose estimate was proposed. Here, with the aim of assessing the generality of the method, the general formulation is particularized for a given differential driven robot. Its kinematic model relates two internal measurements: the instantaneous displacement of both, right and left wheels. The obtained formulation is validated experimentally and compared against Kalman filtering.