Improved odometry motion models for differential-like mobile robots

Flores I.A.C. · 2020 IEEE ANDESCON, ANDESCON 2020 (2020) · DOI: 10.1109/andescon50619.2020.9272088

Resumen

To describe the motion of mobile robots we use probabilistic models that take into account the uncertainty inherent in the motion. This work presents an improved odometry model to describe the motion of mobile robots with differential-like drives. The approach presents an improved algorithm to compute the odometry controls which can handle better the case of in-place rotation and backwards motion. It also presents a new odometry model in which the covariance matrix is only dependent of the amount of motion performed by the robot, and the case of in-place rotation is described by a Gaussian approximation. The main advantages of this model are its invariance to the threshold values used to update it in the prediction step of filter algorithms, a better description of the case of in-place motion, and its capability to handle backwards motion.

Instituciones: Universidad Nacional de Ingeniería