Improved Tracking and Docking of Industrial Mobile Robots Through UKF Vision-Based Kinematics Calibration

Performing an open-loop movement, or docking, for industrial mobile robot (IMR), is a common necessary procedure when relying on environmental sensors not possible. This precision and outcome, solely depend the IMR forward kinematic odometry correctness, which tied to kinematics parameters, depending kind. Calibrating parameters of time consuming mandatory procedure, since mechanical tolerances assembly may introduce large variation from nominal parameters. Furthermore, calibration inaccuracies might severe inconsistencies in tasks such as localization, mapping, navigation general. In this work, we focus so-called parameter calibration. We propose use unscented Kalman filter perform geometrical platform. The platform externally tracked during phase, using fixed temporary external sensor that retrieves position visual tag filter, phase collected data, estimates enlarged system state, comprised have be estimated, position. method can either used online, identify monitor their value while operating, offline, logged data. validate two different devices, 4 mecanum-wheel IMR, Turtlebot 3, camera track movement trough reference chessboard, then comparing original path its corrected version.