Abstract
Localization is a well-studied problem in the field of mobile robotics and is a challenging task as observations like motion estimates or sensor readings are subject to errors. To accurately estimate a robot's pose, such errors need to be considered and thus modeled. In this work, we focus on estimating a robot's pose after motion commands were executed on it. Therefore, an approach to automatically estimate the parameters of the classical Velocity Motion Model using least squares optimization is proposed. It is assumed that the commanded velocities differ from the actual velocities, as noise is distorting the robot's motion. The proposed approach was tested on artificially generated measurements, samples acquired using a simulated robot, and data acquired by conducting experiments with a real robot. The results show that the approach performs better for the measurements acquired with the real robot than with the samples generated in a simulated environment.
Originalsprache | englisch |
---|---|
Titel | Proceedings of the Austrian Robotics Workshop 2022 |
Untertitel | Robotics for Assistance and in Healthcare |
Seiten | 42-47 |
ISBN (elektronisch) | 978-3-99076-109-0 |
Publikationsstatus | Veröffentlicht - 2022 |
Veranstaltung | Austrian Robotics Workshop 2022: ARW 2022 - Villach, Österreich Dauer: 14 Juni 2022 → 15 Juni 2022 |
Konferenz
Konferenz | Austrian Robotics Workshop 2022 |
---|---|
Kurztitel | ARW 2022 |
Land/Gebiet | Österreich |
Ort | Villach |
Zeitraum | 14/06/22 → 15/06/22 |