Intelligent transportation systems are often required to estimate their ego state globally while matching their trajectory to the motion of an object of interest (OoI). For example, industrial robots need to synchronize their motion to the trajectory of an OoI to perform visual inspection task in motion. For that, the ego state has to be estimated both globally and relatively to the OoI. In this work, a pose graph method is proposed, which holistically solves the global and relative state estimation. We demonstrate through closed-loop software-in-the-loop simulation that under ideal noise parameter settings, the proposed approach provides comparable performance to that of a Kalman filter, which is proven to be optimal under such conditions. Furthermore, it is shown that the proposed approach is more robust to inaccurate noise parameters than a Kalman filter. Finally, it is shown by experimental validation that the overall procedure performs effectively on a real robot.
Pose Graph Fusion for Robot Follow Control
2021-09-19
1146642 byte
Conference paper
Electronic Resource
English
Graph-Optimization base multi-sensor fusion for robust UAV pose estimation
BASE | 2017
|AN EXPERIMENTAL STUDY ON RELATIVE AND ABSOLUTE POSE GRAPH FUSION FOR VEHICLE LOCALIZATION
British Library Conference Proceedings | 2018
|