A grasping plan is an algorithm used by a robot arm and gripper to get hold of a specific object. The details of such a program depend on the type of robot arm and gripper, on the size of the robot's work area, and on the type of object(s) to be grasped.
As an example, imagine you own a mobile, autonomous personal robot. You are having spaghetti with lots of sauce for supper, so you feel compelled to instruct the machine to get you six paper napkins. The robot must use a certain program to find the kitchen, another program to locate the cabinet or drawer in which the napkins are stored, and another program to determine which objects in the drawer are the paper napkins. Then the gripper must pick up exactly six napkins without upsetting the other contents of the drawer or cabinet. This is a simple process for a human being, even a child, to carry out, but it is complex when reduced to digital components. There are myriad ways in which the robot might execute this assignment wrong, but only one way to do it right.
A grasping plan can be enhanced by machine vision . Tactile sensors, pressure sensors, and texture sensors can also help ensure that the gripper picks up the correct object in the right way. An efficient and reliable grasping plan involves the use of a detailed, three-dimensional map of the robot's entire work environment, stored in the robot controller's memory. The grasping plan can then assign every object or set of objects a range of coordinate values.