Flexible Vision-Based Auto-docking Control System for Unmanned Ground Vehicles Equipped with Differential Chassis

This paper proposes configurable, parametrically flexible, vision-based automatic docking control system for unmanned surface vehicles (mobile robots) equipped with a differential chassis. The proposed system uses only fiducial localization source to find and attach the robot to docking station. It exploits peculiarities of multiple tags setups to obtain final push-in docking motion straight with high geometrical precision. The trajectories with inverse curvatures are implemented to decrease the number of docking attempts with respect to the actual orientation of the charging slot.