Inchworm robot grippers in clothes manipulation — optimizing the tracing algorithm

Abstract

Unlike factory robots that handle rigid objects most of the time, home service robots need to be able to handle soft objects like clothes, rubber, paper etc. Dexterous manipulation is important when dealing with these deformable objects. For example, in clothes manipulation, it is important for the robot to recognize the shape of the clothes to identify it before sorting or folding it. In order to do that, the robot must first reveal the original state of the clothes which can be done by spreading it. Procedure for clothes spreading for any clothes will always involve grasping of two correct corners before the item is spread. Finding a corner of the clothes can be done by using image processing. The second corner is relatively harder to find because it has to be a corner next to the first corner and not appose it. Edge tracing method ensures the corner found to be a corner next to the first corner. One difficulty during edge tracing of clothes by a robot is to make it trace smoothly and fast without dropping the clothes in the process. The problem is how to retrieve or regrasp the clothes when it is in danger of slipping away from the gripper. This task is difficult due to the fact that deformable objects are sensitive to contact forces. One solution to this problem is to design a gripper that can trace without having to fear for the clothes to slip away. The usage of inchworm type grippers is proposed. Two sets of grippers will allow the robot to trace the clothes more freely because either one gripper will always be holding the clothes. This paper discusses the optimization of the parameters for edge tracing using the inchworm type grippers for towel spreading through experiment.