In a factory or warehouse, if a mechanical part (e.g., a small gear) is accidentally left on the floor — or a robot dropped an item while operating — it may end up in the path of another robot (e.g., autonomous mobile robot, or AMR, delivering items). The robot may not be able to detect the object on the floor, which could cause the robot to drive over it, possibly getting the object caught in its mobile base. This could cause the robot’s wheels to lock up, halting warehouse operations, and possibly cause the robot to fall, especially if the center of gravity is high.
In a more populated retail store environment, customers may drop small items (e.g., trash, pens, wallets) on the floor which is shared by service robots (e.g., Softbank Robotics Pepper, deployed in Microsoft Stores: https://www.youtube.com/watch?v=v931a6sPUZw), posing similar issues to the warehouse environment, but with a higher potential of harming humans (i.e., if the robot falls, it could hit an employee or customer).
In a home environment, similar issues to the previous two scenarios are posed, with additional concerns for the presence of children, who are less alert to the danger of small objects on the floor. A toddler routinely drops and throws small items, posing dangers for themselves and others, including robots.
We propose that a mobile manipulator robot equipped with a 3D LiDAR would be able to detect these objects on the floor such that they could avoid collisions with the objects and move the object out of the way to limit further safety hazards for humans and other robots.