A New Concept of Safety Affordance Map for Robots Object Manipulation

When a robot picks an object, the robot needs to search a safe spot to place it. The safe spot is usually determined away from robots falling, dropping, or hitting an object. Searching of such safe spots may also be necessarily found, when a sharp object like a knife is beside kids. In oder to search a safe area for placing an object, this paper introduces a new concept of safety affordance map (SAM), in which a service robot with an arm is able to manipulate an object in safety. The safety affordance map (SAM) represents a safe area to avoid high risk that may happen in the robot's operational space. The accidnet is the falling down or dropping off of an object by a unskilled agent like a child or an old person. SAM is built by considering environment information as shape of table or objects mutual relationship like a cup of water and a keyboard beside it. SAM could be able to provide appropriate precautions for a robot to avoid an accident. The proposed concept of SAM was examined through several situational experiments in virtual environments.