Towards Task-Speciﬁc Modular Gripper Fingers Automatic Production of Fingertip Mechanics Johannes Ringwald Samuel Schneider Lingyun Chen Dennis Knobbe Lars Johannsmeier

2025-05-06 0 0 5.82MB 8 页 10玖币

侵权投诉

Towards Task-Speciﬁc Modular Gripper Fingers: Automatic Production

of Fingertip Mechanics

Johannes Ringwald, Samuel Schneider, Lingyun Chen, Dennis Knobbe, Lars Johannsmeier,

Abdalla Swikir and Sami Haddadin

Abstract— The number of sequential tasks a single gripper

can perform is signiﬁcantly limited by its design. In many cases,

changing the gripper ﬁngers is required to successfully conduct

multiple consecutive tasks. For this reason, several robotic tool

change systems have been introduced that allow an automatic

changing of the entire end-effector. However, many situations

require only the modiﬁcation or the change of the ﬁngertip,

making the exchange of the entire gripper uneconomic. In

this paper, we introduce a paradigm for automatic task-speciﬁc

ﬁngertip production. The setup used in the proposed framework

consists of a production and task execution unit, containing

a robotic manipulator, and two 3D printers - autonomously

producing the gripper ﬁngers. It also consists of a second

manipulator that uses a quick-exchange mechanism to pick

up the printed ﬁngertips and evaluates gripping performance.

The setup is experimentally validated by conducting automatic

production of three different ﬁngertips and executing grasp-

stability tests as well as multiple pick- and insertion tasks,

with and without position offsets - using these ﬁngertips. The

proposed paradigm, indeed, goes beyond ﬁngertip production

and serves as a foundation for a fully automatic ﬁngertip design,

production and application pipeline - potentially improving

manufacturing ﬂexibility and representing a new production

paradigm: tactile 3D manufacturing.

I. INTRODUCTION

Most conditions of manipulation tasks in industrial pro-

duction setups can be precisely controlled, like initial posi-

tions and orientations of manipulation objects which should

be processed. Despite these controllable conditions, per-

forming tasks with multi-degree of freedom (DoF) hands is

still a complex assignment and a subject of basic research.

Moreover, the performance of soft grippers depends on the

scenario considered. For example, robustness problems can

arise if objects with sharp edges must be grasped with

high grasping forces. Accordingly, most industrial assembly

This work was funded by the German Research Foundation (DFG,

Deutsche Forschungsgemeinschaft) as part of Germany’s Excellence Strat-

egy – EXC 2050/1 – Project ID 390696704 – Cluster of Excellence

“Centre for Tactile Internet with Human-in-the-Loop” (CeTI) of Technische

Universit¨

at Dresden. The authors gratefully acknowledge the funding of the

Lighthouse Initiative KI.FABRIK Bayern by StMWi Bayern (KI.FABRIK

Bayern Phase 1: Aufbau Infrastruktur and KI.Fabrik Bayern Forschungs-

und Entwicklungsprojekt, grant no. DIK0249). We also gratefully acknowl-

edge the funding of the Lighthouse Initiative Geriatronics by LongLeif

GaPa gGmbH (Project Y).The authors are with the Chair of Robotics and

Systems Intelligence and Munich Institute of Robotics and Machine Intel-

ligence, Technical University Munich (TUM), D-80797 Munich, Germany

e-mail: {firstname.surname}@tum.de

?Please note that S. Haddadin has a potential conﬂict of interest as a

shareholder of Franka Emika GmbH.

?This work has been submitted to the IEEE for possible publication.

no longer be accessible.

Fig. 1. Automatic ﬁngertip production and task execution based on a

given ﬁngertip design and ﬁnger base set - using two robot-arms and two

3D printers.

tasks are still performed by two- or three-ﬁnger parallel

grippers. To enable robust grasping and manipulation of

deﬁned objects, the ﬁngers of these grippers must be man-

ually designed, produced, tested and iterated until reaching

a satisfactory performance. This adaptation process is not

only very laborious and time-consuming, but also requires a

high level of design experience in this ﬁeld [1]. Therefore,

the number of assembly objects and manipulation scenarios

that can be handled is limited by the time a designer needs

to adapt a ﬁnger to a speciﬁc application. This makes the

assembly line setup in-ﬂexible, since any changes, like tool-

changing, ﬁnger adaption etc. are very time consuming

and therefore costly. Accordingly, assembly or production

lines are usually setup for a few product types, which are

produced in masses over a longer time period (month/years).

All integrated robots conduct only one speciﬁc task with

one specialized tool or ﬁnger design, which increases the

required number of robots for a designated assembly scenario

signiﬁcantly.

A ﬂexible production of smaller batch sizes requires a

smaller adaption time of the setup to a new product (min-

utes/hours). Therefore, a signiﬁcantly shorter - ideally fully

automatized - gripper ﬁnger design, production and iteration

process would be needed.

Several automatic approaches for robot ﬁnger design adap-

tation have been introduced to solve the aforementioned

problems [2]–[9]. One aspect of this adaptation process

concerns the ﬁnger surface design which majorly affects the

ability to robustly perform stable grasps and manipulation

tasks. Different approaches have been introduced to optimize

the gripper ﬁnger morphology and geometry based on a

given set of objects [1], [10]–[16]. Despite all the efforts to

automate the gripper ﬁnger design, the set of tasks a robot

manipulator can perform is still limited to its mounted ﬁn-

gers. Therefore, different tool-changing systems for robotic

applications have been developed [17]–[32]. While most of

arXiv:2210.10015v1 [cs.RO] 18 Oct 2022

Fig. 2. Presented testbed divided into a production and a task-execution unit. The production unit conducts the extension of 3D printed standard ﬁnger-bases

with custom ﬁngertip elements by: 1. extraction of two ﬁnger-bases from corresponding ﬁnger-base magazines by a robot-arm. 2. Insertion and locking of

these ﬁnger-bases to ﬁnger-holder frames, stationary placed on the build-plate of two 3D printers. 3. Printing of custom ﬁngertips on top of pre-produced

ﬁnger-bases. 4. The extended ﬁnger is picked up by the robot-arm and inserted into the magazines of the quick-ﬁnger-exchange system. 5. Using the QFE

system, the gripper ﬁngers are picked up by the task execution unit. To conduct the ﬁnger testing tasks an IoT-Box is used, which provides pick and place

tasks for three different objects: key, Ethernet-cable, battery.

these systems focus on exchanging the entire end-effector,

some of them focus speciﬁcally on replacing single ﬁngers

[17], [20], [31], or modifying the ﬁnger conﬁguration [32].

Unfortunately, none of the aforementioned work provides

a fully automated process for the gripper ﬁnger production

and exchange. In other words, the ﬁngers must be replaced

manually during an ongoing manipulation process.

Motivated by the above limitations, we introduce a fully

automated setup for gripper ﬁnger production and exchange.

In particular, by integrating a collaborative robot - the Panda

arm - with two 3D printers, it was possible to automate the

production of a predeﬁned set of ﬁngertips. Furthermore, we

introduce a passive quick-ﬁnger-exchange system that allows

the robot to automatically change the used ﬁnger-pair during

a manipulation task. By combining the automatic ﬁngertip

production and the quick-ﬁnger-exchange system, the time

required to adapt to a new assembly task dramatically

decreases. E.g., the automatic production of new ﬁngertips

using a standard ﬁnger base reduces the production time of

new gripper ﬁngers signiﬁcantly (minutes instead of hour).

The number of assembly steps a single robot can perform

during a new production/assembly task increases, since one

robot can select and change its gripper-ﬁnger automatically

during the entire assembly process. Accordingly our ap-

proach makes it possible to realize an efﬁcient and ﬂexible

small batch production/assembly process, while minimizing

the required time, labor and infrastructure required for the

assembly - like the number of robots needed for the task.

Fig.1 illustrates this automatic process. Moreover, the pro-

posed approach will serve as a ﬁrst step and foundation for a

fully automated task-speciﬁc ﬁnger design, production, and

application pipeline, which will enable a robot to identify its

own ﬁngertips and, potentially, provide the base for a novel

production paradigm: tactile 3D manufacturing.

The paper is organized as follows. In Sec. II, we describe

the production processes as well as the quick-ﬁnger-exchange

mechanism. Then, in Sec. III, we demonstrate the appli-

cation and effectiveness of the proposed approach through

experiments. Finally, we discuss our results and provide a

conclusion in Sec. IV.

II. PRODUCTION AND TASK EXECUTION UNIT

The proposed setup is divided into a production unit and a

task execution unit, see Fig.2. The production unit performs

all the required steps to automatically build a task-speciﬁc

parallel gripper ﬁngertip design onto a pre-deﬁned ﬁnger-

base. The task execution unit uses these ﬁngers to conduct

a deﬁned set of tasks. Those ﬁngers can be automatically

changed during the robot operation using a quick-exchange

system. Accordingly, the ﬁngers the production unit provides

can be collected to a ﬁnger library and used multiple times

by the task-execution unit - depending on the currently

needed ﬁngertip geometry. This allows the same robots of

an assembly line/setup to change their ﬁngers automatically

in seconds and adapt to a new product quickly.

Therefor, our speciﬁcally designed 3D-printing based pro-

duction setup has the following advantages compared to a

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

TowardsTask-SpecicModularGripperFingers:AutomaticProductionofFingertipMechanicsJohannesRingwald,SamuelSchneider,LingyunChen,DennisKnobbe,LarsJohannsmeier,AbdallaSwikirandSamiHaddadinAbstractThenumberofsequentialtasksasinglegrippercanperformissignicantlylimitedbyitsdesign.Inmanycases,changingthegr...

展开>> 收起<<

Towards Task-Speciﬁc Modular Gripper Fingers Automatic Production of Fingertip Mechanics Johannes Ringwald Samuel Schneider Lingyun Chen Dennis Knobbe Lars Johannsmeier.pdf

共8页,预览2页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

Towards Task-Speciﬁc Modular Gripper Fingers Automatic Production of Fingertip Mechanics Johannes Ringwald Samuel Schneider Lingyun Chen Dennis Knobbe Lars Johannsmeier

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: