Legged robots have received much attention recently due to their maneuverability in various environments. For autonomously operating the legged robots in harsh conditions like bushes, mountains, and mud, stable and accurate state estimation is necessary. However, estimating the robot’s pose under such slippery and visually not informative conditions is hard. Therefore, this paper proposes a novel framework for the legged robot’s state estimation. Our proposed algorithm utilizes the deep inertial-joint factor based on the inertial-joint network. With real-world experiments, our proposed system validates that the deep inertial-joint factor can help improve the performance of the state estimator.
Inertial-Joint Learning-Aided Robust Pose Estimation for Legged Robots
Lect. Notes in Networks, Syst.
International Conference on Robot Intelligence Technology and Applications ; 2023 ; Taicang December 06, 2023 - December 08, 2023
22.11.2024
7 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Vision-Aided Inertial Navigation for Pose Estimation of Aerial Vehicles
British Library Conference Proceedings | 2009
|Legged robots and methods for controlling legged robots
Europäisches Patentamt | 2023
|LEGGED ROBOTS AND METHODS FOR CONTROLLING LEGGED ROBOTS
Europäisches Patentamt | 2019
|