A Multi-Curvature Soft Gripper Based on Segmented Variable Stiffness Structure Inspired by Snake Scales

Abstract

In atypical industrial settings, soft grippers needed to adjust to different object shapes. Existing grabbers typically accommodated only single-curvature, fixed-stiffness objects, restricting their stability and usability. This study presents a design for a finger featuring multi-curvature, incorporating a wedge actuator alongside two variable stiffness units (VSUs) inspired by snake scales. By adjusting the high stiffness and low stiffness states of the variable stiffness element, the local structural stiffness of the finger was changed, thereby granting the gripper capabilities in bending shape control and variable stiffness. A finite element model of the wedge actuator was developed, and the influence of several parameters, including top wall thickness, side wall thickness, transition layer thickness, and sidewall height on bending angle and tip output force was analyzed through an orthogonal experiment. Furthermore, the relationship between the longitudinal length of the wedge actuator and both the bending angle and the tip output force was studied. Via explicit dynamic analysis, the stiffness variation of the VSU under operational vacuum pressure was predicted and subsequently validated against experimental data, confirming the reliability of the model. The effectiveness of finger shape control and stiffness adjustment was evaluated through experiments. Ultimately, a two-finger gripper was constructed to carry out the grasping experiments. The results showed that the gripper is capable of generating various clamping curvatures, enabling it to conform closely to the objects it grips and significantly broaden its clamping range.

Introduction

With the increasing requirements for adaptability and work diversity, the typical robotic working environment has recently shifted from a single working cell to an unstructured working environment. In many tasks, determining how to interact reasonably with working objects to achieve better execution has emerged as a fundamental research problem in robotics.^1,2 Although rigid grippers offer precise control and high stiffness, their complex control and poor interaction make them ill-suited for unstructured environments. For adaptability and safety, soft robots made of soft materials have been rapidly developed.^3,4 Grasping, a common operation in soft robots, allows soft grippers to attach non-destructively to fragile or multi-curvature object surfaces, demonstrating strong interactivity. Research on soft grippers is also developing rapidly.^5,6 Currently, soft grippers can be categorized by actuation methods into cable-driven,^7,8 shape memory alloy-driven,^9–11 magnetic field-driven,^12–14 and pneumatic-driven soft grippers,^15–18 among which pneumatic drive, due to its simplicity, high output force, low-cost, and mainstream status, has seen significant development.

Despite advancements, most pneumatic soft grippers only achieve single-curvature bending, making it challenging to perfectly fit objects of different sizes. Their broader application still requires additional technical advancements, such as multi-curvature configuration. A multi-curvature structure is a necessary condition for a soft gripper to adapt to the shape and size of the object it grasps through segmented bending characteristics, thereby enhancing its performance. Inspired by human dense arrays, Jeong et al.¹⁹ divided the pneumatic driving cavity and combined it with a variable stiffness mechanism based on particle interference for multi-curvature configuration. Yang et al.²⁰ designed a substrate with three-layer filling and two-layer particle filling cavities, utilizing layer and particle interference differences, and selectively altered regional stiffness for multi-curvature. Al-Rubaiai et al.²¹ proposed a compact, low-cost stiffness adjustment mechanism based on conductive polylactic acid material, demonstrating stiffness and deformation control in soft robotics with soft pneumatic actuators. Wang et al.²² introduced a soft claw based on shape memory alloy for the first time, changing hinge stiffness and driving the shape memory alloy wire for multiple postures. Yoshida et al.²³ proposed a pneumatic bending actuator integrated with a variable stiffness of low melting point alloy for multi-point bending and shape retention. These studies achieved multi-curvature configuration through layer and particle-blocking variable stiffness structures or shape memory alloys. However, there are few reports about the bionic structure realizing the stiffness change and integrating the wedge actuator to achieve the multi-curvature shape of a soft finger. Sun et al.²⁴ developed a variable stiffness gripper featuring a stiffness-adjustable layer inspired by pangolin scales, demonstrating a promising approach. Their findings establish the potential of bionic structures to adjust the local stiffness of the soft gripper.

In this study, a multi-curvature soft gripper based on a segmented variable stiffness structure inspired by snake scales was designed. Firstly, the soft gripper combined with a wedge actuator and two variable stiffness units (VSUs) inspired by snake scales was designed and prepared. On this basis, the effects of structural geometric parameters on the bending performance and end output force of the wedge actuator are further analyzed by orthogonal experimental design method and finite element simulation to optimize the structural design of the soft gripper. Then, the finite element model of the VSU was established based on its bionic design and working principle, and the variable stiffness characteristics were analyzed and verified by experiment. Finally, the influence of the stiffness of VSU on the bending stiffness and multi-curvature behavior of the soft gripper was analyzed, and the adaptability and gripping force of the soft gripper were assessed through a series of experiments tests.

Design and Manufacture

The gripper is composed of two soft fingers, as shown in Figure 1a. These fingers comprise a wedge actuator and two VSUs. The soft finger undergoes bending deformation when the wedge actuator is engaged. By applying vacuum pressure, the stiffness of the VSU can be altered between high and low states. This adjustment allows for the modification of the local structural stiffness of the soft finger, enabling it to achieve various curvature configurations, as depicted in Figure 1b.

FIG. 1.

Schematic diagram of soft gripper: (a) Double-finger gripper; (b) VSU states. VSUs, variable stiffness units.

Design of the wedge actuators

As shown in Figure 2a, the wedge actuator has a width of 38 mm and a wedge angle α of 75°. Other parameters are discussed as follows. To enhance the bending performance and tip output force of the wedge actuator, a wedge-shaped soft cavity structure was designed, as shown in Figure 2b. The thickness of each wall in the soft cavity structure, made of silicone rubber (Dragon 30, Smooth-On), varies. A strain-limiting layer, composed of fiberglass mesh and silicone rubber, is bonded to the bottom of the cavity to prevent longitudinal elongation of the base. In contrast, the driving layer remains capable of elongation. Incorporating pre-stressed glass fiber mesh in the confined layer helps mitigate the “balloon effect” at the bottom plate due to driving air pressure.²⁵ Consequently, with the application of appropriate driving air pressure, bending deformation becomes the predominant motion.

FIG. 2.

Design of wedge actuator: (a) Diagrammatic diagram; (b) cross-section; and (c) cross-sectional parameters of the soft chamber.

In this study, we adopted an orthogonal experimental design method to develop a simulation model scheme comprising four factors, each at three levels, resulting in a total of nine groups. The primary parameters of the soft cavity design (Xi) include the side wall thickness Ta (2 ± 0.5 mm), the top wall thickness Tb (3.8 ± 0.5 mm), the side wall height Hb (10 ± 5 mm), and the connecting wall thickness Hc (3 ± 1 mm) (Fig. 2c). The orthogonal experimental design method optimizes combinations involving multiple factors, minimizes the number of trials, and enables a comprehensive analysis of the effects of various factors on the results. The finite element model for each group was developed under a driving pressure of 20 kPa, the results are shown in Table 1.

Table 1.

Bend Angle and Output Force Results from Orthogonal Experiments

Number	Hc	Hb	Ta	Tb	Bend angle (°)	Output force (N)
1	1	1	1	1	73.32	0.483
2	1	2	3	2	70.75	0.625
3	1	3	2	3	80.12	1.695
4	2	1	3	3	54.22	0.522
5	2	2	2	1	71.82	2.208
6	2	3	1	2	81.94	2.416
7	3	1	2	2	56.36	1.224
8	3	2	1	3	68.19	2.438
9	3	3	3	1	66.51	2.437

In the orthogonal experimental design method, selecting the optimal parameter combination involves analyzing the factor level sum K_i (sum of test results at each level), factor level mean k_i (average of K_i, with the maximum indicating the optimal level), and extreme difference R_i (difference between maximum and minimum k_i values). A larger R_i suggests a greater influence of the factor. As shown in Table 2, R (Hb) is the largest for both the bending angle and end output force, indicating the significant influence of Hb. For the bending angle, although k₁ (Hc) is highest, the influence of Hc remains limited. Conversely, k₃ (Hc) is highest for end output force. Therefore, the Hc = 4 mm is selected. From the results of bending angle and end output force, k₃ (Hb), k₁ (Ta), and k₁ (Tb) are optimal, and hence Hb = 15 mm, Ta = 1.5 mm, and Tb = 3.3 mm are selected.

Table 2.

Bend Angle and Output Force Range Analysis

	Bend angle (degree)				Output force (N)
	Hc (mm)	Hb (mm)	Ta (mm)	Tb (mm)	Hc (mm)	Hb (mm)	Ta (mm)	Tb (mm)
K ₁	224.19	183.90	223.45	211.65	2.80	2.23	5.34	5.13
K ₂	207.98	210.76	208.30	209.05	5.15	5.27	5.13	4.27
K ₃	191.06	228.57	191.48	202.53	6.10	6.55	3.58	4.66
k ₁	74.73	61.30	74.48	70.55	0.93	0.74	1.78	1.71
k ₂	69.33	70.25	69.43	69.68	1.72	1.76	1.71	1.42
k ₃	63.69	76.19	63.83	67.51	2.03	2.18	1.19	1.55
R	11.04	14.89	10.66	3.04	1.10	1.44	0.58	0.16

The calculation formulas for the K_i value, the average value result k_i, and the variance R_i are as follows: $K_{i = 1, 2, 3} (X) = \sum X_{i = 1, 2, 3}$ (1) $k_{i = 1, 2, 3} (X) = K_{i = 1, 2, 3} (X) / 3$ (2) $R_{i = 1, 2, 3} (X) = \max k_{i = 1, 2, 3} (X) - \min k_{i = 1, 2, 3} (X)$ (3)

ABAQUS was used to simulate the bending angle and tip output force of actuators with varying lengths. The bending model (Fig. 3a) involved specifying material parameters, using static analysis steps, constraining one end, and applying pressure to the chamber wall to produce a bending effect. Based on this model, the output force model (Fig. 3b) added an analytic rigid body aligned and fixed to the far center of the actuator. A reference point was set in the remote center of the actuator to generate an output force curve. The driving pressure of the actuator in the model is 20 kPa. To characterize the material properties of silicone rubber, a uniaxial tensile test was carried out according to the GB/T528-2009. Four tensile specimens were tested at a rate of 50 mm/min, the specimens were stretched to fracture, and stress-strain curves were acquired. The Yeoh model is applied to the stress-strain data, allowing the determination of material parameters through a uniaxial tensile test.²⁶ The corresponding strain energy function is expressed as: $W = \sum_{i = 1}^{N} C_{i} {(I_{1} - 3)}^{i} + \sum_{k = 1}^{N} \frac{1}{d_{k}} {(J - 1)}^{2 k}$ (4)

FIG. 3.

Finite element (FE) of actuator: (a) Bending angle under different longitudinal lengths; (b) output force model; and (c) variation of bending angle and output force.

For incompressible materials, it can be simplified to: $W = \sum_{i = 1}^{N} C_{i} {(I_{1} - 3)}^{i}$ (5)

Let n = 3, to obtain the third-order expansion of the Yeoh model: $\begin{array}{l} σ = 2 C_{1} (6 ε^{5} - 5 ε^{4} + 4 ε^{3} - 3 ε^{2} + 3 ε) \\ + 12 C_{2} (8 ε^{5} - 5 ε^{4} + 3 ε^{3}) + 54 C_{3} (3 ε^{5} - ε - 1) \end{array}$ (6)

By importing the stress-strain data into the Yeoh model with n = 3 in the finite element analysis, C₁ = 0.153, C₂ = 0.046, C₃ = 0.001 were obtained.

The soft finger incorporating a variable stiffness element and a wedge actuator, prioritizes multi-curvature performance. Through the orthogonal experiment design method, we modified the transition regions between cavities by 1 mm increments, analyzing the wedge actuator’s length at 7 mm intervals. As shown in Figure 3c, the maximum force and maximum bending angle were observed at 128 mm and 135 mm, respectively. However, the selection of these two actuator lengths leads to the problems of too short a longitudinal length of the actuator and too high a structural height of the VSU, which causes the two VSUs to interfere during the bending process of the soft gripper. Therefore, a suitable longitudinal length needs to be selected to ensure the normal bending of the soft gripper while ensuring sufficient end output force and bending angle. For the longitudinal length beyond 142 mm, both performance metrics declined. For below 142 mm, the variable stiffness element obstructs the bending process of the soft gripper due to its structural height. Thus, 142 mm was selected as the optimal actuator length.

The wedge actuator proposed exhibits several advantages over conventional rectangular actuators.^15,27,28 First, the introduction of a top fillet design enhances the tip force output. Second, the design avoided direct contact between the pneumatic mesh and the bottom, reduced the constraint of the bottom, and increased the bending angle and output force by 20% and 40% respectively. Third, the actuator is designed with a thicker top and connecting walls, and thinner side walls to achieve optimal bending angles and tip output forces. Finally, the wedge aerodynamic mesh offers superior negative curvature compared to the rectangular aerodynamic mesh’s contour restriction.

Design of VSU inspired by snake scales

Biological studies,^29–31 reveal that the abdominal scales of most snakes overlap, featuring oval-shaped small grooves on their outer surface and protrusions on the inner surface. One end of each abdominal scale is anchored in the subcutaneous tissue. Muscle movements alter the angle between the scales and the snake’s body, allowing the upper and lower scales to interlock and thereby stiffen the body. Inspired by these characteristics of snake scales, a VSU was designed, as shown in Figure 4a. This unit comprises three artificial scales, three connecting rods, and a rubber base. A rotary joint connects each artificial scale to the soft base, facilitating changes in the angle between them. The surfaces of the artificial scales feature a toothed structure, with the front teeth being slightly larger than the rear ones to prevent jamming.

FIG. 4.

Design and experiments of variable stiffness units: (a) Bionic snake scales structure; (b) experimental device; (c) force-displacement curves under different air pressure; and (d) comparison of experimental and numerical results of force-displacement in high stiffness and low stiffness.

The outer skin acts as a part of the sealing mechanism and contracts when air is evacuated, causing the dentate structures on adjacent artificial scales to interlock through a jamming principle, thus increasing structural stiffness. Introducing positive pressure inflates the outer skin back to its original state, disengaging the artificial scales and transitioning the structure to a low-stiffness state. The stiffness of the assembly can be altered by controlling the air pressure, allowing for stiffness modulation through simple air pumping without the need for additional equipment. This ensures straightforward operation and high efficiency.

As shown in Figure 4b, the experimental apparatus was utilized to determine the optimal vacuum pressure and VSU stiffness tests. Install a lifting platform under the tensile force meter to adjust the height. One end of the VSU is fixed on the sliding guide rail, and the other end is in contact with the push tension meter, and the corresponding displacement is moved to record the reading of the push tension meter. The force-displacement diagram of VSU was obtained by repeating the above experimental process under four different vacuum pressures, illustrated in Figure 4c. The diagram indicates that the variable stiffness performance is inadequate at lower vacuum pressures. On the contrary, at a working pressure of −60 kPa, the skin connection was prone to cracking, so −50 kPa was selected as the working pressure.

The simulation of the VSU involved specifying material parameters, utilizing the display dynamics analysis step, and completely fixing one end of the VSU. The transition between high and low stiffness states was achieved by applying negative pressure to the inner cavity. Self-contact settings were used to prevent mold penetration, resulting in the final data output. The related material parameters are detailed in Table 3. The variable stiffness performance of the VSU was evaluated. Stiffness, defined by the slope of the force-displacement curve, was calculated to determine the stiffness of the VSU. As presented in Figure 4d, on the basis of the selected vacuum working pressure, the force-displacement diagram of VSU before and after the stiffness change is further measured. According to Figure 4d, the load of VSU increases gradually with the movement of the horizontal platform. The stiffness of the VSU was determined by linear fitting in the displacement range of 0–2.5 mm, with a high stiffness of 0.13 N/mm and a low stiffness of 0.03, resulting in a high-to-low stiffness ratio that is $\frac{0 .13 N/mm}{0 .03 N/mm} \approx 4.3$ .

Table 3.

Material Parameters Used in the Simulation

Property	Young’s modulus (mpa)	Poisson’s ratio	Density (g/cm³)
White resin	2000	0.3	1.1
Nylon	1000	0.35	1
Rubber	300	0.45	1

Fabrication of wedge actuator and VSU

The multi-curvature finger proposed in this study comprises a wedge actuator and a snake-like scale VSU. The integration of the wedge actuator and the silicone outer skin of the VSU involves solidification through a 3D printing mold, while the artificial scales, connecting rods, and bases are fabricated using 3D printing techniques.

Silicone rubber underwent stirring and vacuuming as a pre-treatment to minimize bubble formation, which is essential for the curing process of silicone rubber. The main structure of the wedge actuator, utilizing silicone rubber, was created by curing in a white resin mold. It was then bonded with the glass fiber mesh and a silicone rubber base plate to construct the wedge actuator (Fig. 5a). In a 3D printed white resin mold, a rectangular silicone rubber coating was produced through curing. Components such as artificial scales (made of black nylon), bases (constructed from styrene-butadiene rubber 40°), and connecting rods (fabricated from white resin) were assembled into the main structure of the snake-scale VSU. This assembly was then placed within the outer skin made of silicone rubber. To fabricate a VSU, a rectangular silicone rubber film (thickness of 1 mm) was encapsulated. It has a length Lv of 54 mm, a width Wv of 38 mm, and a height Hv of 20.5 mm (Fig. 5b). As shown in Figure 5c, the wedge-shaped actuator was joined to VSUs to assemble a multi-curvature soft finger.

FIG. 5.

Fabrication process of the soft finger: (a) A wedge actuator; (b) a VSU; and (c) assembly of VSUs and a wedge actuator.

Result and Discussion

In this section, the stiffness variation ability of the finger and its multi-curvature characteristics were assessed through experiments. In addition, a two-finger gripper was developed to further evaluate the grasping ability of the proposed finger.

Soft finger stiffness test

This experiment evaluates the stiffness performance of the finger. A digital force meter (range: 0–50 N, resolution: 0.001 N) was mounted on a vertically adjustable lifting platform, which could also move horizontally via a guide rail. During testing, the force meter probe was positioned in contact with the fingertip. Finger stiffness was measured at three bending angles: θ = 0°, 45°, and 90° (Fig. 6a). The wedge actuator driving pressure was set to P₁, and the VSU vacuum pressure to P₂. A linear fit of the 0–6 mm displacement range was used to calculate the slope of the force-displacement curve, representing finger stiffness.

FIG. 6.

Stiffness experiments: (a) Schematic of the stiffness tests; (b) force-displacement curve at 0° position; (c) force-displacement curve at 45° position; and (d) force-displacement curve at 90° position; (e) comparison of stiffness at different position and different vacuum pressure levels.

At θ = 0°, stiffness increased from 0.017 N/mm at P₂ = 0 kPa to 0.024 N/mm at P₂ = −50 kPa, with a flexural stiffness ratio of 1.41 (Fig. 6b). At θ = 45°, the flexural stiffness ratio between 0 and −50 kPa was $\frac{0 .088 N/mm}{0 .028 N/mm} \approx 3 .14$ (Fig. 6c), while at θ = 90°, the flexural stiffness ratio was $\frac{0 .098 N/mm}{0 .055 N/mm} \approx 1. 78$ (Fig. 6d). The main reason for this phenomenon is that the change in stiffness of the VSU is constant at the specified vacuum pressure, whereas the stiffness of the actuator gradually increases with the increase in driving air pressure, thus reducing the effect of the VSU on the stiffness of the soft fingers. As illustrated in Figures 6b–6d, the larger the vacuum pressure P₂, the steeper the force-displacement curve, indicating a higher stiffness. To better illustrate this, histograms of stiffness changes under varying bending angle and vacuum pressures were plotted (Fig. 6e), it demonstrates that the influence of VSU on finger stiffness, where the stiffness decreases with bending angle increases. This is due to the higher driving pressure of the actuator, and the greater bending stiffness.

Multi-curvature bending shape test

Most existing soft actuators lack shape control capabilities, yet this ability is crucial for performing a wide range of tasks. In this study, the shape control ability of the proposed soft actuator was achieved by pumping air and selectively controlling the stiffness of the VSU to counteract specific bending moments. To validate the shape control ability of the proposed soft actuator, multi-curvature test experiments were conducted. Figure 7a shows the pneumatic control experiment platform and its operating principle, which supplies the necessary positive air pressure to the finger. Figure 7b depicts the experimental setup for the multi-curvature test. The multi-curvature finger was affixed to a metal platform using a 3D printing fixture. The air pressure of the wedge actuator (P₁) was regulated by a pneumatic proportional valve and a control valve, while the vacuum pressure of the VSU (P₂) was managed through a control valve. The midpoint of the fingertip served as the coordinate, recording the finger’s final position and capturing the motion of the multi-curvature finger with a high-definition camera.

FIG. 7.

Bending experiments: (a) Pneumatic control platform and working principle; (b) schematic diagram; (c) finger bending configurations when driven pressure is 15 kPa; and (d) finger bending configurations when driven pressure is 20 kPa.

The wedge actuators were inflated with 15 kPa and 20 kPa as shown in Figure 7c and Figure 7d, and the VSU was adjusted as follows: In case 1, VSU 1 and VSU 2 were not inflated; In case 2, P₂ of VSU 2 was set to −50 kPa; In case 3, P₂ of VSU 1 was set to −50 kPa. As shown in the figure, we can selectively adjust the stiffness variation of the VSU, which makes the soft finger produce three curvature shapes.

As shown in Figure 7c and Figure 7d, in condition 1, because VSU 1 and VSU 2 were in a low stiffness state, the bending impact on the soft actuator was minimal, resulting in a relatively smooth curvature transition of the main body. In this scenario, the finger exhibited enhanced compliance, making it better suited for securely grasping large or spherical objects. In condition 2, with VSU 2 in a high stiffness state, there was a significant change in the curvature of the main body. In specific areas, the bending moment was countered by the VSU, resulting in nearly zero curvature. This deformation allowed for more secure grasping of small or rectangular objects. Conversely, in condition 3, with VSU 1 in a high stiffness state, the curvature of the main gripper altered minimally, and the latter half exhibited a relatively smooth curvature transition. This configuration was particularly effective for securely grasping pentagonal objects. In summary, by selectively altering the stiffness of the VSU, the multi-curvature finger could adapt to various grasping tasks.

Adaptive grasping test

Initially, a vacuum of 50 kPa was applied to the lower VSU. Subsequently, under pressures of 30 kPa, 40 kPa, and 50 kPa, the multi-curvature gripper and the conventional gripper introduced in this study were employed to grasp a pentagonal object (Fig. 8). Observations from Figure 8 reveal that under identical air pressure, the multi-curvature gripper demonstrated a notably superior effect on the surface of the object compared to the conventional gripper. This superior performance can be attributed to the bending control capability of the multi-curvature gripper. At the same time, the multi-curvature gripper developed in this study exhibited effective encapsulation capabilities for objects with triangular and round hammer shapes.

FIG. 8.

Adaptive grasping experiments: (a) The proposed gripper; (b) the conventional gripper.

Gasping tests of various objects

The VSU stiffness in each finger can be adjusted independently, enabling a variety of configurations to suit the shape of the object being handled. To assess the grasping performance of the double-finger gripper, specific tests were carried out. Successful grasping was defined as the ability to hold the object continuously for four attempts and maintain stability for more than 5 s. A selection of 12 common objects was captured (Fig. 9), showcasing different modes of grasping, such as grabbing and pinching. The results indicate that the gripper’s ability to grasp is directly related to the driving pressure. VSU activation was unnecessary for normal grasping. When grasping heavier objects, all VSUs could be activated. The gripper could successfully grasp objects of various shapes, weighing between 6 and 234.6 grams.

FIG. 9.

Grasping performance experiments of various objects: (a) A paper cup; (b) a double-sided tape; (c) a bottle cap resin product; (d) a resin product; (e) a brush; (f) a stainless steel pipe; (g) a knife; (h) a thermometer; and (i) a tape.

Conclusions

Due to the uncontrollable curvature of most flexible pneumatic grippers, their application has been limited. In this study, a snake-scale-inspired VSU was designed and integrated with a wedge actuator to develop a soft finger capable of bending shape control and variable stiffness, thereby expanding the application scope of the gripper. Finite element simulations were conducted to predict the stiffness variation capability of the VSU and to compare these predictions with experimental results. The findings indicate that at an air pressure of 50 kPa, the high-to-low stiffness ratio of the VSU is 4.3, with an error margin between the experiment results and the simulation predictions of less than 15%. The stiffness variation capability of the soft finger was quantified and assessed through experiments, and the bending shape control capability was evaluated. Finally, a two-finger gripper was constructed, and a demonstration of a grasping task was performed. The results show that, besides grasping common objects, the gripper exhibits an improved fit on shapes such as pentagons, achieving robust grasping. In future work, we aim to integrate soft sensors into the soft fingers, enabling feedback on bending angles for precise shape control, thereby enhancing grasp performance.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This research was supported by the National Natural Science Foundation of China (Grant Nos. 52205134, 52075492, 52405137), the Zhejiang Provincial Natural Science Foundation of China (Grant Nos. ZCLZ24E0502, LD22E050009, LLSSZ24E050002, LQ22E050013, and LGG22E050043), Postdoctoral Preferred Funding Project of Zhejiang Province (ZJ2024130), Taizhou Natural Science Foundation Project (Grant No. 23gya23).

References

Severinson-Eklundh

, Green

, Hüttenrauch

. Social and collaborative aspects of interaction with a service robot. Robot Autonomous Syst, 2003; 42(3–4):223–234; doi: 10.1016/S0921-8890(02)00377-9

Kulic

, Croft

. Affective state estimation for human-robot interaction. IEEE Trans Robot, 2007; 23(5):991–1000; doi: 10.1109/TRO.2007.904899

Fang

, Zhuang

, Liu

, et al. A shift from efficiency to adaptability: Recent progress in biomimetic interactive soft robotics in wet environments. Adv Sci (Weinh), 2022; 9(8):e2104347; doi: 10.1002/advs.202104347

Jing

, Qiao

, Pan

, et al. An overview of the configuration and manipulation of soft robotics for on-orbit servicing. Sci China-Inform Sci, 2017; 60(5); doi: 10.1007/s11432-016-9033-0

Shintake

, Cacucciolo

, Floreano

, et al. Soft robotic grippers. Adv Mater, 2018; 30(29):e1707035; doi: 10.1002/adma.201707035

Hainsworth

, Smith

, Alexander

, et al. A fabrication free, 3D printed, multi-material, self-sensing soft actuator. IEEE Robot Autom Lett, 2020; 5(3):4118–4125; doi: 10.1109/LRA.2020.2986760

Fang

, Sun

, Wu

, et al. Multimode grasping soft gripper achieved by layer jamming structure and tendon-driven mechanism. Soft Robot, 2022; 9(2):233–249; doi: 10.1089/soro.2020.0065

Giorelli

, Renda

, Calisti

, et al. Neural network and jacobian method for solving the inverse statics of a cable-driven soft arm with nonconstant curvature. IEEE Trans Robot, 2015; 31(4):823–834; doi: 10.1109/TRO.2015.2428511

Yang

, Xu

, Li

, et al. Design and implementation of a soft robotic arm driven by SMA coils. IEEE Trans Ind Electron, 2019; 66(8):6108–6116; doi: 10.1109/TIE.2018.2872005

10.

Patterson

, Sabelhaus

, Majidi

. Robust control of a multi-axis shape memory alloy-driven soft manipulator. IEEE Robot Autom Lett, 2022; 7(2):2210–2217; doi: 10.1109/LRA.2022.3143256

11.

She

, Chen

, Shi

, et al. Modeling and validation of a novel bending actuator for soft robotics applications. Soft Robotics, 2016; 3(2):71–81; doi: 10.1089/soro.2015.0022

12.

Ebrahimi

, Bi

, Cappelleri

, et al. Magnetic actuation methods in bio/soft robotics. Adv Funct Materials, 2021; 31(11); doi: 10.1002/adfm.202005137

13.

Marullo

, Salvietti

, Prattichizzo

. On the use of magnets to robustify the motion control of soft hands. IEEE Robot Autom Lett, 2022; 7(4):11743–11750; doi: 10.1109/LRA.2022.3205751

14.

Ding

, Zhou

, Sun

, et al. Magnetorheological elastomer actuated multi-stable gripper reinforced stiffness with twisted and coiled polymer. Thin-Walled Structures, 2023; 193:111223; doi: 10.1016/j.tws.2023.111223

15.

Guo

, Elgeneidy

, Xiang

, et al. Soft pneumatic grippers embedded with stretchable electroadhesion. Smart Mater Struct, 2018; 27(5):55006; doi: 10.1088/1361-665X/aab579

16.

Fatahillah

, Oh

, Rodrigue

. A novel soft bending actuator using combined positive and negative pressures. Front Bioeng Biotechnol, 2020; 8:472; doi: 10.3389/fbioe.2020.00472

17.

Zhang

, Ni

, Wu

, et al. Pneumatically actuated soft gripper with bistable structures. Soft Robot, 2022; 9(1):57–71; doi: 10.1089/soro.2019.0195

18.

Zhang

, Ni

, Gao

, et al. Pneumatically controlled reconfigurable bistable bionic flower for robotic gripper. Soft Robot, 2022; 9(4):657–668; doi: 10.1089/soro.2020.0200

19.

Song

, Lee

, Moon

, et al. A multi-curvature, variable stiffness soft gripper for enhanced grasping operations. Actuators, 2021; 10(12):316; doi: 10.3390/act10120316

20.

Yang

, Zhang

, Kan

, et al. Hybrid jamming for bioinspired soft robotic fingers. Soft Robot, 2020; 7(3):292–308; doi: 10.1089/soro.2019.0093

21.

Al-Rubaiai

, Pinto

, Qian

, et al. Soft actuators with stiffness and shape modulation using 3D-printed conductive polylactic acid material. Soft Robot, 2019; 6(3):318–332; doi: 10.1089/soro.2018.0056

22.

Wang

, Ahn

. Shape memory alloy-based soft gripper with variable stiffness for compliant and effective grasping. Soft Robot, 2017; 4(4):379–389; doi: 10.1089/soro.2016.0081

23.

Yoshida

, Morimoto

, Zheng

, et al. Multipoint bending and shape retention of a pneumatic bending actuator by a variable stiffness endoskeleton. Soft Robot, 2018; 5(6):718–725; doi: 10.1089/soro.2017.0145

24.

Sun

, Chen

, Han

, et al. A soft gripper with variable stiffness inspired by pangolin scales, toothed pneumatic actuator and autonomous controller. Roboti Computer-Integrated Manufa, 2020; 61:101848; doi: 10.1016/j.rcim.2019.101848

25.

Tian

, Zhang

, Yuan

, et al. Research and analysis of stiffness-enhanced soft gripper. Ifs, 2021; 40(4):6231–6238; doi: 10.3233/JIFS-189460

26.

Marechal

, Balland

, Lindenroth

, et al. Toward a commom framework and database of materials for soft robotics. Soft Robot, 2021; 8(3):284–297; doi: 10.1089/soro.2019.0115

27.

Hao

, Wang

, Ren

, et al. Modeling and experiments of a soft robotic gripper in amphibious environments. Inter J Adv Robotic Systems, 2017; 14(3):172988141770714; doi: 10.1177/1729881417707148

28.

Mosadegh

, Polygerinos

, Keplinger

, et al. Pneumatic networks for soft robotics that actuate rapidly. Adv Funct Materials, 2014; 24(15):2163–2170; doi: 10.1002/adfm.201303288

29.

Liu

, Chen

, Zheng

, et al. Wear resistance improvement of keeled structure and overlapped distribution of snake scales. J Bionic Eng, 2023; 20(3):1121–1131; doi: 10.1007/s42235-022-00300-2

30.

Zhao

, Su

, Hou

, et al; Department of Electrical and Computer Engineering, National University of Singapore, 4 Engineering Drive 3, 117576, Singapore. Directional sliding of water: Biomimetic snake scale surfaces. Oea, 2021; 4(4):210008–210008; doi: 10.29026/oea.2021.210008

31.

Spinner

, Kovalev

, Gorb

, et al. Snake velvet black: Hierarchical micro- and nanostructure enhances dark colouration in bitis rhinoceros. Sci Rep, 2013; 3:1846; doi: 10.1038/srep01846