Design and fabrication of a soft robotic gripperfor object picking

 CÔNG NGHỆ 
 Tạp chí KHOA HỌC VÀ CÔNG NGHỆ ● Tập 57 - Số 2 (4/2021) Website: https://tapchikhcn.haui.edu.vn 78
KHOA HỌC P-ISSN 1859-3585 E-ISSN 2615-9619 
DESIGN AND FABRICATION OF A SOFT ROBOTIC 
GRIPPERFOR OBJECT PICKING 
THIẾT KẾ VÀ CHẾ TẠO TAY KẸP MỀM ROBOT DÙNG CHO GẮP VẬT 
Hoang Hong Hai 
ABSTRACT 
In this paper, author proposed a model of a soft robotic gripper made of 
elastic rubber material that can pick up a variety of objects. The gripper consists 
of three soft fingers evenly spaced on the truncated conical surface. The finger is 
made by molding method and design of the finger is inherited from "Pneunets 
bending actuator" of soft robotics toolkit. This paper shows soft finger model size 
parameters, finite element simulation is performed to validate finger design. 
Also, the soft finger fabrication process is introduced together with the driving 
circuit and control circuit to perform the soft gripper opening - closing operation. 
Experiments are carried out to assess the finger's elasticity at different pressures 
and pressures required for handling objects of different sizes and materials. The 
experiment is shown that the soft gripper design could be remarkable applied to 
pick and place robot for soft object. 
Keywords: Soft robotic, grasping, image processing, gripper. 
TÓM TẮT 
Trong bài báo này, tác giả đề xuất một mô hình tay kẹp rô bốt mềm được 
làm bằng vật liệu cao su đàn hồi có thể gắp được đồ vật khác nhau. Tay kẹp bao 
gồm ba ngón mềm đặt đều nhau trên bề mặt hình nón cụt, ngón được làm bằng 
phương pháp đúc và thiết kế của ngón được thừa hưởng từ "Pneunets bending 
actuator" của bộ công cụ robot mềm. Bài báo đưa ra các thông số kích thước mô 
hình ngón tay mềm, mô phỏng phần tử hữu hạn được thực hiện để xác nhận 
thiết kế ngón tay. Ngoài ra, quy trình chế tạo ngón tay mềm được giới thiệu cùng 
với mạch dẫn động và mạch điều khiển để thực hiện thao tác đóng - mở kẹp gắp 
mềm. Các thí nghiệm được thực hiện để đánh giá độ đàn hồi của ngón tay ở các 
áp suất và lực kẹp khác nhau cần thiết để xử lý các vật có kích thước và vật liệu 
khác nhau. Thí nghiệm chỉ ra rằng, thiết kế tay kẹp mềm có thể ứng dụng tốt cho 
robot gắp vật dạng mềm. 
Từ khóa: Robot mềm, gắp vật, xử lý ảnh, tay kẹp. 
School of Mechanical Engineering, Hanoi University of Science and Technology 
Email: hai.hoanghong@hust.edu.vn 
Received: 01/3/2021 
Revised: 05/4/2021 
Accepted: 25/4/2021 
1. INTRODUCTION 
In the increasingly developed society, the food industry 
plays an important role in life. Each large-scale production 
plants which can market millions of consumer goods each 
year in a variety of sizes and volumes. Currently, the 
packaging of products manufacture line is still handled by 
humans due to the possibility of product distortion or 
damage. In order to minimize costs, an automated 
packaging system is required. The most commonly used 
automatic packaging system today is a rigid gripper and 
vacuum sensor. However, hard gripper can damage food 
ingredients and the suction cup requires a flat surface. 
Recently, soft robotic grippers can be categorized into 
three groups based on grasping principle [1]. To lift the 
object, the gripper is used base on soft pneumatic 
actuators [2, 3] the MR fluid gripper [4], and the jamming 
gripper [5] or even to pick up thin object the soft gripper is 
used suction [6, 7], the electrostatic gripper [8], and 
grippers based on gecko adhesion [9, 10] 
In this study, author proposed the design and 
fabrication with a soft gripper based on the referenced 
research and design toolkit [11-15]. The gripper design is 
introduced in Section 2, and simulations with control are 
shown in Section 3. Experimental validations are presented 
in Section 4. Section 5 presents our conclusions. The 
experimentally pick up some soft objects for evaluation 
and shows the good result which could be applied in real 
time application. 
2. DESIGN 
The finger design and fabrication are based on Pneunet 
and soft robotics toolkit. However, for simplicity in 
fabrication, we have removed the "stress-limiting layer" in 
Pneunet's finger (a rectangular paper placed inside the 
finger). As shown in the Figure 1, the finger is the same size 
as the Asian man, consisting of 12 gas chambers, the 
distance between two consecutive chambers is 1.5mm and 
the largest chamber has a thickness of 3mm, which makes 
the end of the finger to be harder to mimic the function of 
human nails. The 1.2mm high air path is designed to pass 
through all compartments allowing air to pass through all 
chambers. A 4mm diameter hole at the top allows for the 
mounting of air ducts from the outside to the fingers. Wavy 
structure is designed at the surface under the finger to 
increase friction when in contact with the object and easier 
to hold. 
In order to connect the fingers together and form a 
complete gripper, author designed the connector and base 
for the soft hand. Figure 2a and Figure 2b show the lower 
and upper halves of the terminals. During assembly, the 
P-ISSN 1859-3585 E-ISSN 2615-9619 SCIENCE - TECHNOLOGY 
Website: https://tapchikhcn.haui.edu.vn Vol. 57 - No. 2 (Apr 2021) ● Journal of SCIENCE & TECHNOLOGY 79
finger is inserted into the lower half of the connector, then 
thread the air tube through the upper and lower half of the 
connector respectively, finally fixing the connector to the 
finger with a bolt and nut. The height of the compartment 
created by the two halves of the connector is designed to 
be 1mm shorter than the height of the finger, so the finger 
will be fixed after assembling the connector. In Figure 2c, 
author design a clamp base for the soft gripper with three 
hole connectors that are evenly spaced conical. The quick 
locking mechanism between the two shaft couplings 
(lower half of the connector) and the hole (clamp base) 
makes fitting the connector with the clamp base more 
convenient. The complete assembly model in Figure 2d 
with 3 soft fingers, we choose 3 fingers because this is the 
minimum number of fingers to stabilize an object, the 
stability increases if author add the number of soft fingers 
in gripper. Materials used for fabrication are ABS plastic and 
FDM 3D printing technology because it ensures fabrication 
accuracy and detailed durability. 
Figure 1. Design of the soft finger 
Figure 2. a) Bottom connector, b) base, c) upper connector d) 3D gripper 
model 
Figure 3. 3D printed molds 
Figure 4. Materials for finger fabrication 
a) 
b) 
Figure 5. Soft grippers after fabricating and assembling 
 CÔNG NGHỆ 
 Tạp chí KHOA HỌC VÀ CÔNG NGHỆ ● Tập 57 - Số 2 (4/2021) Website: https://tapchikhcn.haui.edu.vn 80
KHOA HỌC P-ISSN 1859-3585 E-ISSN 2615-9619 
The soft finger is fabricated by traditional casting 
method with Dragon Skill 10 silicon fabrication material 
(Smooth-on Inc., PA, USA). This material is used extensively 
to make masks and as molds for the manufacture of 
defective parts that need to be replaced. The mold as 
shown in Figure 5b is printed by using the Anycubic Mega 
S 3D printer. The tools and materials required for the 
molding process are shown in Figure 6. 
Figure 6. Finger fabrication process 
Final soft gripper product after fabricating and 
assembling is shown in Figure 5. Figure 6 describes the soft 
finger fabrication process. The number of fingers is three 
because this is the minimum number for stable handling 
and increased grip stability with increased number of 
fingers. To make soft fingers, mix the silicon solution well 
and de-foam the first time for 2 minutes in a vacuum 
chamber at a pressure of about 85kPa. 
After defoaming, pour the silicon mixture into the lower 
mold of the body and the finger base. After filling the two 
molds with silicone solution, degassing a second time. 
Next, fit the top and bottom molds of the body and finger 
base together and fix them with bolts. After the silicone has 
solidified, carefully demoulding and obtaining two 
separate finger parts. To stick these two parts together, 
place the body on top of the finger base and use the same 
silicone solution to bond them. 
3. SIMULATION AND CONTROL 
3.1. Simulation 
The finite element method is a relatively accurate 
method for testing the deformation ability of an object in 
the mechanical system, commonly used to model and 
simulate soft robots. In this article, author used the 
software Abaqus 2017 (SIMULIA, Dassault System, MA) to 
simulate soft fingers. Dragon Skin 10 material (DS10, 
Smooth-on Inc., PA) is used for the soft finger pattern. 
According to the article [1], DS10 can be characterized as a 
hyper-elastic rubber using the Yeoh model and has 
coefficients such as: C10 = 3.6×10-2MPa, C20 = 2.58×10-4MPa, 
C30 = -5.6×10-7MPa. PTHH model is meshed according to 
tetrahedron element (C3H10) and approximate mesh size is 
2.5mm. Weight is also included in the simulation and the 
density of the material is 1070kg/m3. Pressure loads are 
applied to all of the soft finger inner cavities and surface 
interactions are also generated between adjacent surfaces. 
Standard simulations were applied and used nonlinear 
stress-strain curves to account for large bending strains. 
Figure 7. The soft finger shrinks at a pressure of 10kPa 
Figure 8. Soft fingers stretch at pressures of 5kPa, 10kPa, and 17kPa 
Figure 8 depicts the finger shape at different applied 
pressures. Finger tension simulations show a gradual 
increase in curvature from top to bottom. At a pressure of 
17 kPa, the resulting tension angle α is 72.1 degrees. 
When a negative pressure is applied, the finger will 
shrink and bend in the opposite direction. Simulating 
contraction of the soft finger, the results converge at a 
pressure of -10kPa (Figure 7), where the finger cavities are 
completely closed. The inversion angle (β in Figure 7) is 
obtained by 15.6 degrees, which is calculated by pixel 
coordinates in Paint. 
3.2. Control 
A mini 8HP air compressor is used for finger air supply. 
Because fingers only need less than 100kPa of air pressure, 
P-ISSN 1859-3585 E-ISSN 2615-9619 SCIENCE - TECHNOLOGY 
Website: https://tapchikhcn.haui.edu.vn Vol. 57 - No. 2 (Apr 2021) ● Journal of SCIENCE & TECHNOLOGY 81
most compressors will do. The system uses two 2/2 
solenoid valves controlled by two relays that receive signals 
from arduino. The control interface is designed by 
Windows Form quite simply to serve the functions of 
grabbing, releasing and resting. 
Figure 9. Control diagram 
Two solenoid valves use a 24V source and are 
connected to systems such as Figure 9. After installation we 
have a real model and the control interface as shown in 
Figure 10 and 11. 
Figure 10. Electrical system 
Figure 11. Console control interface 
Valve status in each mode of soft gripper is shown in 
Table 1. 
Table 1. Status of valves at different modes 
Mode Valve 1 Valve 2 
Pick ON OFF 
Drop OFF ON 
Rest OFF OFF 
4. EXPERIMENT RESULTS 
The design robot arm and real model is described in 
Figure 12. Experimental process assessing the force when 
handling and lifting objects with different volumes and 
weights as shown in Figure 13. Initially the gripper will 
open wide and lower down to the place where the object 
is, then the gripper will lift the object from the surface of 
10cm. The calibrated with difference objects size as shown 
in Table 2. 
Table 2. Results of grabbing difference objects size 
Object Pressure[kPa] Mass[g] Size(LxWxH)[mm] 
Bread 20 90 55x55x65 
Egg 20 36 55x44x23 
Orange 20 100 60x60x40 
a) 
b) 
Figure 12. Design model and real model 
Figure 13. Experimental pick-up 
 CÔNG NGHỆ 
 Tạp chí KHOA HỌC VÀ CÔNG NGHỆ ● Tập 57 - Số 2 (4/2021) Website: https://tapchikhcn.haui.edu.vn 82
KHOA HỌC P-ISSN 1859-3585 E-ISSN 2615-9619 
The experiment shows that the clamp can lift objects 
with the size of about 100g with the size of 70x70x70mm. 
With objects with a high friction surface, the ability to grasp 
will be higher, with objects with smooth surfaces like 
tomatoes, the clamps are difficult to lift accurately. 
5. CONCLUSION 
The soft gripper crafted can only lift light objects, not 
yet able to lift fruits with a smooth surface and high 
density. To improve the ability of soft grippers, it is possible 
to design the finger tips more undulating to create more 
friction. Grasping tests also showed that the gripper can 
adapt to different target shapes and sizes by adjusting the 
initial opening and finger configurations. The gripper 
successfully lifted a small, ellipsoidal egg (L55mm × 
W55mm × H65mm), circular orange (L60mm × W60mm × 
H60mm) in the perpendicular configuration. At the same 
time the finger can be well crafted to withstand high air 
pressure and handle even more heavy objects as weight as 
1kg. The proposed design of soft gripper show that it can 
facilitate automation process in food manufacturing 
process. 
REFERENCES 
[1]. Zhongkui Wang, Keung Or, Shinichi Hirai, 2020. A dual-mode soft gripper 
for food packaging. Robotics and Autonomous Systems. 
[2]. K. Suzumori, S. Iikura, H. Tanaka, 1991. Development of flexible 
microactuator and its applications to robotic mechanisms. Proceedings of the 1991 
IEEE international Conference on Robotics and Automation, pp. 1622–1627. 
[3]. Y. Chen, S. Guo, C. Li, H. Yang, L. Hao, 2018. Size recognition and 
adaptive grasping using an integration of actuating and sensing soft pneumatic 
gripper. Robot. Auton. Syst. 104, 14–24. 
[4]. A. Pettersson, S. Davis, J.O. Gray, T.J. Dodd, T. Ohlsson, 2010. Design of a 
magnetorheological robot gripper for handling of delicate food products with 
varying shapes. J. Food Eng. 98, 332–338 
[5]. E. Brown, N. Rodenberg, J. Amend, A. Mozeika, E. Steltz, M.R. Zakin, 
H.Lipson, H.M. Jaeger, 2010. Universal robotic gripper based on the jamming of 
granular material. Proceedings of the National Academy of Sciences of the 
United States of America, pp. 18809–18814. 
[6]. S. Davis, J.O. Gray, D.G. Caldwell, 2008. An end effector based on the 
Bernoulli principle for handling sliced fruit and vegetables. Robot. Comput.-Integr. 
Manuf. 24, 249–257. 
[7]. T.K. Lien, P.G.G. Davis, 2008. A novel gripper for limp materials based on 
lateral Coanda ejectors. CIRP Ann. 57, 33–36 
[8]. E.W. Schaler, D.R. III, P. Glick, V. White, A. Parness, 2017. An electrostatic 
gripper for flexible objects. Proceedings of the 2017 IEEE/RSJ International 
Conference on Intelligent Robots and Systems, pp. 1172–1179 
[9]. E.W. Hawkes, D.L. Christensen, A.K. Han, H. Jiang, M.R. Cutkosky, 2015. 
Grasping without squeezing: shear adhesion gripper with fibrillar thin film. 
Proceedings of the 2015 IEEE International Conference on Robotics and 
Automation, pp. 2305–2312 
[10]. E.W. Hawkes, H. Jiang, M.R. Cutkosky, 2015. Three-dimensional 
dynamic surface grasping with dry adhesion. Int. J. Robot. Res. 35 (8), 943–958 
[11]. https://softroboticstoolkit.com/synergistic-design/fabrication/pneunet 
[12]. Zhongkui Wang, Mingzhu Zhu, Sadao Kawamura, Shinichi Hirai 2017. 
Comparison of diferent soft grippers for lunch box packaging. Robotics and 
Biomimetics volume 4, Article number: 10. 
[13]. https://softroboticstoolkit.com/resources-for-educators/soft-gripper 
[14]. https://docs.opencv.org/master/dc/dbb/tutorial_py_calibration.html 
[15]. C. Cattaneo, G. Mainetti, R. Sala, 2015. The Importance of Camera 
Calibration and distortion Correction to Obtain Measurements with Video 
Surveillance Systems. Journal of Physics: Conference Series 658, 01200. 
THÔNG TIN TÁC GIẢ 
Hoàng Hồng Hải 
Viện Cơ khí, Trường Đại học Bách khoa Hà Nội