Research and design of a multi-fingered hand made of hyperelastic material

Junhui Zhang (School of Civil Engineering, North China University of Technology, Beijing, China)
Xiufeng Zhang (National Research Center for Rehabilitation Technical Aids, Beijing, China)
Yang Li (School of Aerospace Engineering, Tsinghua University, Beijing, China)

Assembly Automation

ISSN: 0144-5154

Publication date: 6 August 2018

Abstract

Purpose

The purpose of this study is to provide a novel multi-fingered hand made of hyperelastic material. This kind of hand has the advantage of less mechanical parts, simpler control system. It can greatly cut down the complexity and cost of the hands under conditions of ensuring enough flexibility of grasping.

Design/methodology/approach

Based on the principle of virtual work, the equations of pulling force and grasping force are derived. To get the max grasping force, the optimal structural dimensions of the hand are obtained by finite element simulations. Hand’s grasping experiment is conducted.

Findings

The factors influencing grasping force and grasping stability are identified, and they are the length between short poles around the knuckles and the height of short poles. Experimental results show that the max strain of knuckles is less than the elastic limit of hyperelastic material, and the presented hand is practicable. The adaptive ability and grasping stability of the presented hand are demonstrated.

Originality/value

A novel multi-fingered hand made of hyperelastic material is presented in this paper. By designing the thickness of every section of a hyperelastic plate, the knuckle sections will bend and other sections of the plate will remain straight, and thus, the multi-fingered hand will grasp.

Keywords

Citation

Zhang, J., Zhang, X. and Li, Y. (2018), "Research and design of a multi-fingered hand made of hyperelastic material", Assembly Automation, Vol. 38 No. 3, pp. 249-258. https://doi.org/10.1108/AA-03-2017-042

Download as .RIS

Publisher

:

Emerald Publishing Limited

Copyright © 2018, Emerald Publishing Limited


1. Introduction

Multi-fingered humanoid hands, which are widely used as general operating equipment in agriculture and industry fields, can also be assembled in prosthetic arms for disabled people. At present, multi-fingered humanoid hands mostly are all-driven multi-fingered dexterous hands such as Shadow Dexterous Hand (Kochan, 2005), KIST Hand (Kim et al., 2011), Robonaut Hand (Bridgwater and Ihrke, 2012) and UBV Hand (Palli and Scarcia, 2012). The dexterous hands can grasp and hold various objects; however, some features such as many actuator elements, complex control system and high cost make them difficult to be widely applied in a short time. Based on design principles of application requirements, low cost and easy operation, underactuated multi-fingered hands have been studied in recent years. Underactuated hands have fewer degrees of actuators (DoA) than degrees of freedom (DoF). Examples of such hands and their intended application area are, for instance, the prosthetic FDD hand (Visser and Herder, 2000) consisting of three actuated fingers and two passive fingers with a total of 5 DoF and 1 DoA; Kamikawa and Maeno (2008) developed a five-fingered prosthetic hand at Keio University (15 DoF, 1 DoA); the SARAH (Laliberté et al., 2002) intended for space applications consists of three fingers (10 DoF, 2 DoA) and the pneumatically driven TWIX hand (Bégoc et al., 2007) was intended for industrial applications (6 DoF, 1 DoA). More examples of this can be found in Birglen et al. (2008). Due to the underactuation of these hands, the fingers intrinsically conform to the shape of the objects. Such hands are therefore able to grasp a variety of objects without controlling the closing motion of the digits. This means that underactuated hands may constitute an affordable yet effective category of grippers suitable for picking up and placing operations with different objects in unstructured environments. The underactuated multi-fingered hands have the advantage of strong grasping, stable holding, less actuators, simpler control and low cost. Owing to these advantages, underactuated multi-fingered hands have dramatically increased the practicability of multi-fingered dexterous hands.

In recent years, compliant or soft underactuated hands have been extensively studied owing to their special characteristics. Giannaccini et al. (2014) presented a soft cable-driven gripper, featuring no stiff sections, which can adapt to a wide range of objects because of its entirely soft structure. Odhner et al. (2014) introduced the iHY Hand, whose fingers are capable of both firm power grasps and low-stiffness fingertip grasps using only the compliant mechanics of the fingers. Sheng et al. proposed a novel underactuated prosthetic finger with a compliant driving mechanism. Li and Qiao (2015) introduced the physical structure and the control mechanism of human motor nervous system to the robotic system in a tentative manner to improve the compliance of the robot. Manti et al. (2015) developed a soft gripper by exploiting the combination of soft materials, underactuated mechanisms and a bioinspired design. Deimel and Brock (2016) presented the RBO Hand 2, a highly compliant, underactuated, robust and dexterous anthropomorphic hand. Qiao et al. (2016) proposed a new motion model based on the human motion pathway, and this proposed motion model proves to have fast response and learning ability through experiments. Li et al. (2016) proposed a novel glove-based virtual hand grasping approach for virtual mechanical assembly, and a prototype system is designed and developed to implement the proposed approach. Stuart et al. (2017) developed a underactuated, compliant, tendon-driven robotic hand, which uses elastic finger joints and a spring transmission to achieve a variety of pinch and wrap grasps. Stavenuiter et al. (2017) introduced the concept design of an underactuated grasper with the ability to adjust its level of self-adaptability by changing the rotational stiffness of its differential mechanism. A bi-stable mechanism was implemented in this hand. More examples of soft robotics can be found in Rus and Tolley (2015).

In this paper, a novel multi-fingered hand is presented and the finger of the hand is made of hyperelastic material. By designing the thickness of every section of a hyperelastic plate, the knuckle sections will bend and other sections of the plate will remain straight; thus, the multi-fingered hand can grasp objects. This kind of hand has the advantage of less mechanical parts, simpler control system. It can greatly cut down the complexity and cost of the hands under conditions of ensuring enough flexibility of grasping.

The paper is structured as follows: First, a literature survey is provided on the underactuated multi-fingered hands. The components and mechanism of a multi-fingered hand made of hyperelastic material are introduced in Section 2. Based on the principle of virtual work, the equations of pulling force and grasping force are derived in Section 3. The optimal structural dimensions of the hand are obtained by finite element simulations in Section 4. The hand’s grasping experiment is conducted in Section 5. The Section 6 is the conclusions.

2. Mechanism of a multi-fingered hand made of hyperelastic material

A finger of a multi-fingered hand is shown in Figure 1, which is constituted by one plate, six short poles and one string. The plate is made of hyperelastic material, thickness of cross-section of the plate is varied and thin sections are used as knuckles. The six short poles are fixed on the plate. The string is threaded through the holes in the short poles; one tip of the string is bonded to the pole at fingertip and the other tip is connected to a small motor. Because of small thickness, bending stiffness of the knuckle sections is smaller than other sections of the plate. When the motor pulls the string, the knuckle sections will bend and other sections of the plate will remain straight, and thus, the multi-fingered hand will grasp. When the motor loosens the string, restoring force of hyperelastic material will turn the knuckle sections back to flat and the multi-fingered hand will release.

In Figure 1, l1 is the length from the finger root to the center of the first knuckle; l2 is the length between the center of the first knuckle and the center of the second knuckle; l3 is the length between the center of the second knuckle and the center of the third knuckle; l4 is the length between the center of the third knuckle and the fingertip; w is the width of the hyperelastic plate; a1 is the length of the first knuckle; a2 is the length of the second knuckle; a3 is the length of the third knuckle; b1 is the length between the first pair of short poles around the first knuckle; b2 is the length between the second pair of short poles around the second knuckle; b3 is the length between the third pair of short poles around the third knuckle; h0 is the thickness of the hyperelastic plate; h1 is the thickness of the first knuckle; h2 is the thickness of the second knuckle; h3 is the thickness of the third knuckle; hc is the distance between the string and the plate.

3. Optimization of a multi-fingered hand made of hyperelastic material

Grasping force and grasping stability are very crucial for a multi-fingered hand. Optimizing the dimensions to obtain the maximum grasping force and grasping stability are presented in this section. Subjected to pulling force T generated by a motor, the knuckles will bend as shown in Figure 2. F1, F2 and F3 are the counterforces acting on the finger due to the hand’s grasping a target object.

Based on the principle of virtual work:

(1) Tv=FTV
where v is the pulling velocity of the string:
(2) v=ṡ
and s is the moving distance of the string:
(3) F=[F1F2F3]
(4) V=[v1v2v3]
v1 is the velocity of acting point of F1; v2 is the velocity of acting point of F2; v3 is the velocity of acting point of F3.

It can be known easily:

(5) [v1v2v3]=[e100f1cosθ2e20f1cos(θ2+θ3)f2cosθ3e3][ω1ω2ω3]
where e1 is the distance between the acting point of F1 and the center of the first knuckle; e2 is the distance between the acting point of F2 and the center of the second knuckle; e3 is the distance between the acting point of F3 and the center of the third knuckle; f1 is the length between the centers of the first and the second knuckles; f2 is the length between the centers of the second and the third knuckles:
(6) [ω1ω2ω3]=[θ̇1θ̇2θ̇3]

Moreover,

(7) ṡ=ṡ1+ṡ2+ṡ3
where s1 is the distance between the tips of the short poles around the first knuckle; s2 is the distance between the tips of the short poles around the second knuckle; s3 is the distance between the tips of the short poles around the third knuckle:
(8) s1=(c1c2cosθ1d2sinθ1)2+(c2sinθ1d2cosθ1d1)2
(9) s2=(c3c4cosθ2d4sinθ2)2+(c4sinθ2d4cosθ2d3)2
(10) s3=(c5c6cosθ3d6sinθ3)2+(c6sinθ3d6cosθ3d5)2

Then:

(11) v1=ṡ1=(c1c2d1d2)sinθ1(c1d2+c2d1)cosθ1(c1c2cosθ1d2sinθ1)2+(c2sinθ1d2cosθ1d1)2×θ̇1
(12) v2=ṡ2=(c3c4d3d4)sinθ2(c3d4+c4d3)cosθ2(c3c4cosθ2d4sinθ2)2+(c4sinθ2d4cosθ2d3)2×θ̇2
(13) v3=ṡ3=(c5c6d5d6)sinθ3(c5d6+c6d5)cosθ3(c5c6cosθ3d6sinθ3)2+(c6sinθ3d6cosθ3d5)2×θ̇3

Then:

(14) v=ṡ=[(c1c2d1d2)sinθ1(c1d2+c2d1)cosθ1(c1c2cosθ1d2sinθ1)2+(c2sinθ1d2cosθ1d1)2(c3c4d3d4)sinθ2(c3d4+c4d3)cosθ2(c3c4cosθ2d4sinθ2)2+(c4sinθ2d4cosθ2d3)2(c5c6d5d6)sinθ3(c5d6+c6d5)cosθ3(c5c6cosθ3d6sinθ3)2+(c6sinθ3d6cosθ3d5)2]T[ω1ω2ω3]

According to equation (1):

(15) [F1F2F3]T[e100f1cosθ2e20f1cos(θ2+θ3)f2cosθ3e3][ω1ω2ω3]=T[(c1c2d1d2)sinθ1(c1d2+c2d1)cosθ1(c1c2cosθ1d2sinθ1)2+(c2sinθ1d2cosθ1d1)2(c3c4d3d4)sinθ2(c3d4+c4d3)cosθ2(c3c4cosθ2d4sinθ2)2+(c4sinθ2d4cosθ2d3)2(c5c6d5d6)sinθ3(c5d6+c6d5)cosθ3(c5c6cosθ3d6sinθ3)2+(c6sinθ3d6cosθ3d5)2]T[ω1ω2ω3]

Then:

(16) [F1F2F3]T=T[(c1c2d1d2)sinθ1(c1d2+c2d1)cosθ1(c1c2cosθ1d2sinθ1)2+(c2sinθ1d2cosθ1d1)2(c3c4d3d4)sinθ2(c3d4+c4d3)cosθ2(c3c4cosθ2d4sinθ2)2+(c4sinθ2d4cosθ2d3)2(c5c6d5d6)sinθ3(c5d6+c6d5)cosθ3(c5c6cosθ3d6sinθ3)2+(c6sinθ3d6cosθ3d5)2]T[e100f1cosθ2e20f1cos(θ2+θ3)f2cosθ3e3]1

Then, the grasping force F3 can be calculated according to equation (16) for a given pulling force T:

(17) F3=Te3(c5c6d5d6)sinθ3(c5d6+c6d5)cosθ3(c5c6cosθ3d6sinθ3)2+(c6sinθ3d6cosθ3d5)2

In equation (17), for a given grasping condition, e3 is invariant and T is known; thus the value of F3 depends on c5, d5, c6 and d6. Based on the Extreme Value Theorem, for the max grasping force, c5, d5, c6 and d6 should satisfy:

(18) {F3c5=0F3d5=0F3c6=0F3d6=0

After c5, d5, c6 and d6 are determined, c1c4 and d1d4 can also be obtained similarly.

4. Grasping simulation

Because of nonlinearity of equation (17), the analytical values of c5, d5, c6 and d6 are difficult to be obtained. In this section, based on the numerical simulations of grasping the optimal c1c6 and d1d6 are determined. The geometrical model and finite element model of a finger created in Abaqus software are shown in Figure 3. The shell element (S4) is used for the finger plate. There are 8,463 nodes and 8,160 shell elements in the model. Material property of the finger plate are shown in Figure 4.

At the top of the short poles, reference points are created; reference points are constrained to the finger plate through multi-point constraints beams. Connector slip rings in interaction module are used to model the string; slip rings connect the reference points of the short poles. The end of the plate is fixed. Constant velocity constraint is applied to the string. The finite element model is solved by the dynamic implicit method.

Figure 5 shows the deformations of a finger in a grasping process. Subjected to the pulling force of the string, the knuckle sections bend and the other sections of the plate remain straight, and the fingers grasp as expected. The maximum mises stress and principal strain of knuckle sections at different time are listed in the Table I. It can be seen that the max stress and strain are in elastic limit of the hyperelastic material, and so a multi-fingered hand made of hyperelastic material is practicable.

Series of numerical simulations have been done to observe the effect of knuckles’ dimensions (c1c6, d1d6) on the pulling force. Figure 6 to Figure 11 show the pulling force with the varying knuckle’s angle for different dimensions of knuckles.

From Figure 6 and Figure 7, it can be seen that pulling forces decrease when d5 and d6 increase. From Figure 8 to Figure 11, it can be seen that pulling forces decrease when c5 and c6 increase. So, the max grasping force can be obtained when the pulling force decreases (i.e. c1c6 and d1d6 increase).

Considering the length limit and thickness limit of a finger, the optimal dimensions are listed in Table II. The whole grasping process simulated in Abaqus is shown in Figure 12. As the pulling force increases, the knuckles bend and the hand grasps gradually. Because of the designing of the thickness of the knuckles, the increasing velocity of the angle of root knuckle and middle knuckle is faster than the end knuckle; therefore, the grasping space and the grasping adaptability are enhanced.

5. Experiment

Considering the experimental purposes and the motion complexity, the thumb is not adopted, and only the forefinger, the middle finger and the ring finger are used in the experiment. Figure 13 shows the designed hyperelastic fingers; from top to bottom, they are the ring finger, middle finger and forefinger. The knuckles are produced by thinning a hyperelastic plate at the specified position to a design thickness. Steel wire passes through the holes in the short poles bonded on the plate. In Figure 13, the steel wire of each finger is attached to a brushless motor, and dynamic strain gauges are installed to monitor the strain of the knuckles.

The control system of the brushless motors is shown in Figure 14. The controller is BeagleBoneBlack, which is a single board computer running Linux operating system. It communicates with personal computer by wireless communication. The type of the motor is Faulhaber3564k024, which is a high-performance brushless and can generate controlled torque by communicating with the controller. Signals of angle, angular velocity and torque are collected by a serial communication board and a digital/analog board. The control system is shown in Figure 15.

The strain of three knuckles of ring finger at different time is listed in Table III. The strain is negative and increases as the knuckles bend. The strain of the middle knuckle is larger than the strain of the root knuckle, and the strain of the end knuckle is the minimum. The max strain of the root knuckle is 4.37 per cent, the max strain of the middle knuckle is 6.53 per cent and the max strain of the end knuckle is 2.97 per cent. They are all less than the elastic limit of hyperelastic material.

The strain of the three knuckles of middle finger at different time is listed in Table IV. The strain is negative and increases as the knuckles bend. The strain of the middle knuckle is larger than the strain of the root knuckle, and the strain of the end knuckle is the minimum. The max strain of root knuckle is 4.46 per cent, the max strain of the middle knuckle is 5.63 per cent and the max strain of the end knuckle is 2.78 per cent. They are all less than the elastic limit of hyperelastic material.

The strain of three knuckles of forefinger at different time is listed in Table V. The strain is negative and increases as the knuckles bend. The strain of the middle knuckle is larger than the strain of the root knuckle, and the strain of the end knuckle is the minimum. The max strain of the root knuckle is 3.31 per cent, the max strain of the middle knuckle is 4.22 per cent and the max strain of the end knuckle is 2.94 per cent. They are all less than the elastic limit of hyperelastic material.

Comparing the results of simulation and experiment shown in Figure 16, it verifies the numerical simulation and the practicability of the multi-fingered hand made of hyperelastic material. By grabbing a cylinder, the adaptive ability and grasping objects ability of the presented hyperelastic multi-fingered hand are demonstrated.

Moreover, the peculiar advantage of this proposed hand is that it has less mechanical parts than other underactuated hands. To compare this hand’s mechanical complexity with other underactuated hands, the number of mechanical parts and the transmission type of some underactuated hands are listed in Table VI. In our design, a finger consists of only one mechanical part (i.e. the hyperelastic plate), which greatly reduces the mechanical complexity of the hand. Thus, the hand’s reliability and controllability can be enhanced.

6. Conclusions

A novel hyperelastic multi-fingered hand is designed in this paper by analysis, calculation, simulation of finite element method and experiment study.

  • A multi-fingered hand made of hyperelastic material is proposed in this paper. This kind of hand has the advantage of less mechanical parts, simpler control system. It can greatly cut down the complexity and cost of the hands under conditions of ensuring enough flexibility of grasping.

  • The equations of pulling force and grasping force are derived. The factors influencing grasping force and grasping stability are identified, and they are the length between short poles around the knuckles and the height of short poles. Optimal structural dimensions of the hand are obtained by finite element simulation.

  • Hand’s grasping experiment is conducted. Experimental results show that the max strain of knuckles is less than the elastic limit of hyperelastic material, and the presented hand is practicable. The adaptive ability and grasping stability of the presented hand are demonstrated.

In a word, the multi-fingered hand made of hyperelastic material proposed in this paper is practicable. This kind of hand has the advantage of less mechanical parts, simpler control system. It can greatly cut down the complexity and cost of the hands under conditions of ensuring enough flexibility of grasping.

Figures

A finger made of hyperelastic material

Figure 1

A finger made of hyperelastic material

Mechanism of the finger

Figure 2

Mechanism of the finger

Finite element model of the finger

Figure 3

Finite element model of the finger

Stress–strain curve of the hyperelastic material

Figure 4

Stress–strain curve of the hyperelastic material

Grasping process of the finger

Figure 5

Grasping process of the finger

Pulling force for various d5 and d6

Figure 6

Pulling force for various d5 and d6

Pulling force for various d5 and d6

Figure 7

Pulling force for various d5 and d6

Pulling force for various c5 and c6

Figure 8

Pulling force for various c5 and c6

Pulling force for various c5 and c6

Figure 9

Pulling force for various c5 and c6

Pulling force for various c5 and c6

Figure 10

Pulling force for various c5 and c6

Pulling force for various c5 and c6

Figure 11

Pulling force for various c5 and c6

Grasping process of the hyperelastic hand

Figure 12

Grasping process of the hyperelastic hand

Hyperelastic fingers

Figure 13

Hyperelastic fingers

Hyperelastic fingers with motors

Figure 14

Hyperelastic fingers with motors

Control system configuration

Figure 15

Control system configuration

Grasping test of the hyperelastic hand

Figure 16

Grasping test of the hyperelastic hand

Stress and strain of the knuckle

t 0.0 3.00 6.00 9.00 12.00 15.00 18.00 21.00 24.00
σ 0.0 1.39 × 102 2.77 × 102 4.11 × 102 5.39 × 102 6.64 × 102 7.84 × 102 9.00 × 102 1.01 × 103
ε 0.0 4.99 × 10−4 7.61 × 10−4 1.31 × 10−3 1.71 × 10−3 1.94 × 10−3 2.14 × 10−3 2.31 × 10−3 2.46 × 10−3

Dimensions of the fingers

\(mm) Ring finger Middle finger Forefinger Thumb
l1 50 50 50 38
l2 42 47 40 36
l3 26 25 24 26
l4 25 25 22 \
a1 10 10 10 10
a2 10 10 10 10
a3 10 10 10 \
b1 26 25 24 36
b2 26 25 24 36
b3 26 25 24 \
h0 2 2 2 2
h1 0.2 0.2 0.2 0.2
h2 0.21 0.21 0.21 0.21
h3 0.25 0.25 0.25 \
hc 10 10 10 10
w 15 15 15 15

Max principal strain of knuckles of the ring finger

Time (s) Root knuckle (%) Middle knuckle (%) End knuckle (%)
1.0 −0.002 −0.002 −0.005
2.0 −0.278 −0.377 −0.529
3.0 −1.364 −1.176 −0.903
4.0 −2.760 −3.963 −1.789
5.0 −3.919 −4.259 −2.113
6.0 −4.517 −4.498 −2.631
7.0 −4.490 −5.441 −2.032
8.0 −4.488 −5.739 −2.972
9.0 −4.286 −6.433 −2.911
10.0 −4.370 −6.531 −2.975

Max principal strain of knuckles of the middle finger

Time (s) Root knuckle (%) Middle knuckle (%) End knuckle (%)
1.0 −0.004 −0.002 −0.006
2.0 −0.033 −0.024 −0.021
3.0 −0.749 −2.269 −1.034
4.0 −1.684 −3.531 −1.076
5.0 −2.502 −3.845 −1.674
6.0 −3.444 −4.199 −2.780
7.0 −3.462 −4.034 −2.734
8.0 −4.032 −4.012 −2.743
9.0 −4.261 −5.917 −2.759
10.0 −4.461 −5.627 −2.786

Max principal strain of knuckles of the forefinger

Time (s) Root knuckle (%) Middle knuckle (%) End knuckle (%)
1.0 −0.006 −0.006 −0.004
2.0 −0.379 −0.290 −0.021
3.0 −0.541 −1.702 −0.029
4.0 −1.023 −2.719 −1.022
5.0 −1.371 −2.305 −1.677
6.0 −1.770 −3.041 −1.855
7.0 −2.080 −3.176 −2.412
8.0 −3.191 −4.193 −2.421
9.0 −3.272 −4.210 −2.795
10.0 −3.309 −4.220 −2.935

Number of mechanical parts of underactuated hands

Hand No. of mechanical parts Transmission type
Hand of this paper 5 Tendon
Hand 1 [14] 10 Air
Hand 2 [12] 12 Tendon
Hand 3 [16] 22 Linkage
Hand 4 [18] 30 Linkage
Hand 5 [6] 35 Tendon

References

Birglen, L., Laliberté, T., Gosselin, C.M., Siciliano, B., Khatib, O. and Groen, F. (2008), “Underactuated robotic hands”, Springer Tracts in Advanced Robotics, Vol. 40.

Bégoc, V., Krut, S., Dombre, E., Durand, C. and Pierrot, F. (2007), “Mechanical design of a new pneumatically driven underactuated hand”, Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), pp. 927-933.

Bridgwater, L.B. and Ihrke, C.A. (2012), “The Robonaut 2 hand-designed to do work with tools”, Proceedings of the 2012 IEEE International Conference on Robotics and Automation, pp. 3425-3430.

Deimel, R. and Brock, O. (2016), “A novel type of compliant and underactuated robotic hand for dexterous grasping”, International Journal of Robotics Research, Vol. 35 Nos 1/3, pp. 161-185.

Giannaccini, M.E., Georgilas, I., Horsfield, I., Peiris, B.H.P.M., Lenz, A., Pipe, A.G. and Dogramadzi, S. (2014), “A variable compliance, soft gripper”, Autonomous Robots, Vol. 36 Nos 1/2, pp. 93-107.

Kamikawa, Y. and Maeno, T. (2008), “Underactuated five-finger prosthetic hand inspired by grasping force distribution of humans”, Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 717-722.

Kim, E.H., Lee, S.W. and Lee, Y.K. (2011), “A dexterous robot hand with a biomimetic mechanism”, International Journal of Precision Engineering and Manufacturing, Vol. 12 No. 2, pp. 227-235.

Kochan, A. (2005), “Shadow delivers first hand”, Industrial Robot: An International Journal, Vol. 32 No No. 1, pp. 15-16.

Laliberté, T., Birglen, L. and Gosselin, C.M. (2002), “Underactuation in robotic grasping hands”, Machine Intelligence and Robotic Control, Vol. 4 No. 3, pp. 1-11.

Li, R. and Qiao, H. (2015), “The compliance of robotic hands - from functionality to mechanism”, Assembly Automation, Vol. 35 No. 3, pp. 281-286.

Li, J.R., Xu, Y.H., Ni, J.L. and Wang, Q.H. (2016), “Glove-based virtual hand grasping for virtual mechanical assembly”, Assembly Automation, Vol. 36 No. 4, pp. 349-361.

Manti, M., Hassan, T., Passetti, G., D’Elia, N., Laschi, C. and Cianchetti, M. (2015), “A bioinspired soft robotic gripper for adaptable and effective grasping”, Soft Robotics, Vol. 2 No. 3, pp. 107-116.

Odhner, L.U., Jentoft, L.P., Claffee, M.R., Corson, S., Tenzer, Y., Ma, R.R., Buehler, M., Kohout, R., Howe, R.D. and Dollar, A.M. (2014), “A compliant, underactuated hand for robust manipulation”, International Journal of Robotics Research, Vol. 33 No. 5, pp. 736-752.

Palli, G.U. and Scarcia, C. (2012), “Development of robotic hands: the UB Hand evolution”, Proceedings of the 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems, Portugal, pp. 5456-5457.

Qiao, H., Li, C., Yin, P.J., Wu, W. and Liu, Z.-Y. (2016), “Human-inspired motion model of upper-limb with fast response and learning ability – a promising direction for robot system and control”, Assembly Automation, Vol. 36 No. 1, pp. 97-107.

Rus, D. and Tolley, M.T. (2015), “Design, fabrication and control of soft robots”, Nature, Vol. 521 No. 7553, pp. 467-475.

Stavenuiter, R.A.J., Birglen, L. and Herder, J.L. (2017), “A planar underactuated grasper with adjustable compliance”, Mechanism and Machine Theory, Vol. 112, pp. 295-306.

Stuart, H., Wang, S., Khatib, O. and Cutkosky, M.R. (2017), “The ocean one hands: an adaptive design for robust marine manipulation”, International Journal of Robotics Research, Vol. 36 No. 2, pp. 150-166.

Visser, H.D. and Herder, J.L. (2000), “Force-directed design of a voluntary closing hand prosthesis”, Journal of Rehabilitation Research and Development, Vol. 37, pp. 261-271.

Supplementary materials

AA_38_3.pdf (27.4 MB)

Acknowledgements

Conflict of interest: The authors declare that there is no conflict of interests.

This project is supported by the project 51335004 of National Natural Science Foundation of China.

Corresponding author

Xiufeng Zhang can be contacted at: zhangxiufeng2863@163.com