A packing problem approach to grasping with a two-finger gripper with application in forest harvesting

Abstract

Robotic Hand Grasping is a challenging task that is studied by different researchers, with contributions made to topics such as grasping quality measuring, synthesis, and mathematical models. However, these works usually consider grasping of only one object at time, which is not applicable when the hand must carry multiple objects. Multiple objects grasping problem can be approached as a packing problem, which consists in filling a container with objects of a determined geometry, using the space available in the most efficient way. The aim of this work is to obtain a design methodology for a two-finger gripper that optimizes the grasping capacity of slender cylindrical multiple objects, applying packing problem concepts. This case is of practical importance in grapple claws used in forest harvesting for picking logs. A mathematical model is presented that allows to obtain an optimized geometric configuration of the grapple claw for maximum grasping capacity and packing density for given numbers and sizes of logs. This model is analytically solved for the maximum radius of a set of logs that can be grasped. Alternatively, for a number of logs with a given diameter, the geometric dimensions of an optimized grapple claw can be obtained. The solutions are evaluated using packing density and Active Force Closure (AFC) as quality metrics, to show the benefits of adding them as design criteria in the design multiple objects grapple claws.