A Mobile Manipulation System for Automated Replenishment in the Field of Unmanned Retail

Automated replenishment is one of the most challenging in-store logistics tasks, which has not been well-addressed due to the complex coordination of a mobile base and a manipulator. In this paper, we build a complete mobile manipulation system to enable automated replenishment in the field of unmanned retail. Specifically, the designed mobile manipulation system consists of a UR5 robotic arm and an automated guided vehicle (AGV). For the manipulation part, we propose a 6-DoF target-oriented grasp detection method for stably picking products in the cluttered scenes; For the mobile part, we develop a navigation method to plan a collision-free path for efficiently delivering the goods. We evaluate the designed system via real-world experiments and the results show the efficiency and reliability of our mobile manipulation system in finishing the replenishment tasks. The video demonstration can be found at https://youtu.be/g6WrF-I7tNw