3D-Printed Origami Actuators for a Multianimal-Inspired Soft Robot with Amphibious Locomotion and Tongue Hunting

Abstract

The field of soft robotics is rapidly evolving, and there is a growing interest in developing soft robots with bioinspired features for use in various applications. This research presented the design and development of 3D-printed origami actuators for a soft robot with amphibious locomotion and tongue hunting capabilities. Two different types of programmable origami actuators were designed and manufactured, namely Z-shaped and twist tower actuators. In addition, two actuator variations were developed based on the Z-shaped actuator, including the pelvic fin and the coiling/uncoiling types. The Z-shaped actuators were used for the rear legs to facilitate the locomotion of the water-like frogs. Meanwhile, the twisted tower actuators were used for the rotation joints in the forelegs and for locomotion on land. The pelvic fin actuator was developed to imitate the land locomotion of the mudskipper, and the coiling/uncoiling actuator was designed for tongue hunting motion. The origami actuators and soft robot prototype were tested through a series of experiments, which showed that the robot was capable of efficiently moving in water and on land and performing tongue hunting motions. Our results demonstrate the effectiveness of these actuators in producing the desired motions and provide insights into the potential of applying 3D-printed origami actuators in the development of soft robots with bioinspired features.

Introduction

Over millions of years, organisms have evolved almost perfect limbs and organs capable of adapting to nature environment. Scholars hope to create highly innovative designs and sophisticated technology in engineering by studying the locomotion of living things. The traditional rigid bioinspired robots have been designed with more mechanical structures,¹ which are more complex and bulkier in a bioinspired sense and do not fit the characteristics of living things. The rise of soft robots in recent years has promoted the development of bioinspired robots, which are endowed with different lifelike characteristics due to the soft material properties and novel actuating technology. Currently, many bioinspired soft robots for use in water are being developed, inspired by jellyfish,² deep-sea snail fish,³ frog,⁴ octopus suckers,⁵ and others, and soft robots inspired by terrestrial reptilian creatures, including earthworm,⁶ crab,⁷ snake,⁸ and lizard.⁹

However, the structures or propulsion methods of these robots can only be adapted to specific environments, regardless of whether they are water or land robots, and an amphibious bioinspired soft robot is challenging to develop.

Amphibians, a class of creatures that have the ability to swim in water and walk on land, offer biological inspiration for a robotic design. Bioinspired soft amphibian robots are always imitating the locomotion and posture of animals.¹⁰ In these studies, the propulsive methods of amphibious robot can be divided into independent and hybrid actuator propulsion methods. Bioinspired robots that use the independent actuator propulsion method are as follows: looper,¹¹ sea urchin,¹² dog,¹³ starfish,¹⁴ tortoise,¹⁵ and lizard,¹⁶ and the propulsion of these robots is similar on land and in water. Although these robots can achieve high efficiency on land, their performance in water is much less than that on land due to the influence of resistance in water. Robots that use hybrid actuator propulsion include EELWORM (inspired by eel and worm),¹⁷ a combination of insect crawling and human swimming postures,¹⁸ a seal,¹⁹ and a combination of turtle and tortoise locomotion postures.²⁰

These robots use multiple actuators and are inspired by multiple biological locomotor modalities. Researchers have taken the locomotor advantages of the various types of organisms and integrated them in one robot, thus greatly increasing the locomotor performance of robots on land and water and enhancing their more integrated locomotor capabilities. However, the problem is that making multiple actuators in different environments without affecting each other is difficult. Specifically, how the robot is actuated in water and on land is difficult to coordinate.

Origami, as an ancient art of paper folding, has received increasing scholarly attention. Origami structures have inspired new approaches to soft robot design,²¹ and many soft origami robots are currently designed based on the following four typical origami structures: Miura origami,^22–24 water bomb origami,^25–27 Yoshimura origami,^28–30 and Kresling origami.^31–33 A complex output can be obtained from a simple input by means of an ingenious origami structure. In fact, the versatility of origami structures combined with the properties of soft materials can bring out the advantages of both, and soft origami can traverse some unstructured areas by folding and assembling itself.³⁴ In previous studies, the modular design of multiple actuators has great advantages for repairing systems, replacing and reconfiguring modules,³⁵ and self-transformation of various shapes and geometries.³⁶ Accordingly, research on the cooperative motion of multiple origami actuators has great potential.

However, despite that single origami structures have been intensively studied, the integration of different patterns of origami actuators is rare.¹⁸ There are fewer examples of research where origami mechanisms may be used in bioinspired robots, and it is underresearched how to combine origami modules with various types to achieve multimodal locomotion of bioinspired soft robots.

This work designed a novel bioinspired amphibian soft robot that integrates the following four origami actuators: Z-shaped, twisted tower, pelvic fin, and coiled/uncoiled origami (Fig. 1A). The motion mechanisms of frog swimming, mudskipper crawling, and frog hunting are observed and analyzed. The result showed that the motion trajectories of the Miura and Kresling origamis could be made to fit the animal's limbs, joints, and organs by a clever design. The four origami actuators are directly 3D printed from a thermoplastic elastomer (TPE) through fused deposition modeling and heat-treated to withstand a high fluid pressure. The motion characteristics of the bioinspired amphibious soft robot can be well characterized with a designed pneumatic hardware control system.

FIG. 1.

Application scenarios and multianimal bioinspiration of the proposed amphibious soft robot. (A) Four origami actuators are integrated into the robot to mimic the frog leg movement to achieve swimming in water, pectoral fin joint of the mudskipper combined with pelvic fin to achieve land crawling, and frog tongue to hunt insects. (B) Proposed robot performing tasks offshore to achieve amphibious locomotion, land–water transition, and hunt pests. Note: Origami I: Z-shaped origami actuator (ZSOA), origami II: twisted tower origami actuator (TTOA), origami III: pelvic fin origami actuator (PFOA), and origami IV: coiling/uncoiling origami actuator (C/UOA).

The robot is actuated by six air paths, which can realize the combined functions of nine different modes (straight/left/right swim, straight crawl/alternate/left/right/back crawl, tongue hunting). The insect catching function can be achieved by relying on the coiled/uncoiled origami actuator (C/UOA) (frog tongue), maximum swimming speed of 0.6 body length per second (BL/s), and turning speed of 10.7°s⁻¹ in water for the robot moving by Z-shaped origami actuators (ZSOAs) (frog legs), combination of twist tower origami actuators (TTOAs) (pectoral fin joint of mudskipper) and pelvic fin origami actuator (PFOA) (pelvic fin of mudskipper), the robot moving with a maximum land crawling speed of up to 20.6 mm/s, and a turning speed of 30° s⁻¹ on land. Meanwhile, the crawling robot can withstand heavy objects of the same weight as itself and can crawl through different terrains, highlighting its ability to move in unstructured environments.

Furthermore, the ability to transition between different environments is an important index for amphibious robots, and a robot with the ability to cross-medium not only from land to water but also from water to land is designed. The application scenario of the robot is also described in Figure 1B.

The main contributions of this article include the following three points: (1) four 3D-printed origami actuators with different morphologies are designed by combining the origami and biological motion mechanisms from two different creatures; (2) programmable motion modes of the origami actuators are evaluated; and (3) the amphibious cross-medium locomotion of a multianimal-inspired soft robot based on origami actuators is realized.

Analysis of Origami Actuators

Structure design and kinematic analysis of ZSOA

Inspiration from Miura origami

The Miura origami method facilitates rapid folding of a paper sheet into multiple origami units, as depicted in Figure 2A. Inspired by Miura origami, we alter the length of the initial origami units to mimic the locomotion of frog leg in water and arranging them to form a Z-shaped structure resembling the tubular configuration proposed by Tachi and Miura³⁷ and used 3D printing technology for the fabrication of ZSOAs. In comparison with Kim et al.'s bioinspired dual-morphing stretchable origami,²² our ZSOA exhibits a simplified fabrication process, and moreover, a frog-like swimming locomotion pattern that is discussed in detail in the following section.

FIG. 2.

Design of ZSOA. (A) The design concept and internal structural parameters of the ZSOA. (B) The motion trajectory of the actuator under varying parameters, indicated by changes in air pressure: Groups 1a, 1b, 1c, and 1d modify channel lengths (c₁, c₂, c₃); Groups 1a, 2a, and 2b vary the length L of the actuator; Groups 1a, 3a, and 3b adjust actuators with width W₁; Groups 1a, 4a, and 4b alter the tilt angle α of the actuator; Groups 1a, 4a, and 4b change wall thickness t₁. (C) Comparison between experimental and simulation data for each group's actuator motion trajectories. Note: During experimentation, the positive pressure P ranged from 0 to 100 kPa, and data at 0, 40, and 100 kPa were sampled and compared.

Actuation mechanism and parameters of ZSOA

The actuator comprises three flexible channels, each capable of being assigned a distinct length. The interior of the actuator consists of a homogeneous gas channel, and programmable motion is achieved by incrementing positive pressure P. The channel is systematically unfolded in accordance with the prescribed design rules, as demonstrated in Supplementary Movie S1, which illustrates the conceptualization of ZSOA. The length of the actuator is denoted as L, the width as W₁, the wall thickness as t₁, the length of the three flexible channels as c₁, c₂, and c₃, and the width tilt angle of the double layer flexible cavity as α. For the parameter configuration of ZSOA, refer to Table S1 in Supplementary Data.

Kinematic analysis of ZSOA

To investigate the kinematics of the actuator in programmable mode, a series of experimental comparisons were conducted under various parameters, as illustrated in Figure 2B. Groups 1a, 1b, and 1c modulate the length of the flexible channel, and with increasing pressure, the two double layer flexible cavities adjacent to the longer channel consistently unfold preferentially. This phenomenon may be attributed to a proportional relationship between the expansion rates of its inner surface. By adjusting the channel lengths, precise control over the motion trajectory of the actuator can be achieved. Groups 1a, 2a, and 2b manipulate the actuator's length, resulting in an increased range of motion that ultimately expands its movement capabilities. The actuator width is modified by Groups 1a, 3a, and 3b, with the degree of deformation being directly proportional to the actuator width.

This relationship may be attributed to the increased cross-sectional stiffness of the material at smaller widths, resulting in reduced elastic deformation when pressure is applied. Groups 1a, 4a, and 4b alter the tilt angle of the actuator, and the magnitude of this tilt angle can impact its deformability within the flow channel. When subjected to identical air pressure conditions, a smaller tilt angle poses greater challenges for achieving substantial deformation in the actuator. The wall thickness of the actuator is altered by Groups 1a, 5a, and 5b. It becomes evident that an actuator with increased wall thickness necessitates higher driving pressure due to the enhanced stiffness resulting from thicker materials, which in turn affects its response speed.

These can be observed in Figure 2C where both experimental and simulated actuator motion trajectories are clearly depicted. Simulation results are depicted in Figure S1 of Supplementary Data. The related experiments are shown in Supplementary Movie S2.

Variation actuators from Z-shaped origami

The Z-shaped origami, inspired by the Miura origami serves as a foundation for the development of various intriguing origami structures. Two types of actuators, the pelvic fin origami and the coiling/uncoiling origami, are derived from Z-shaped origami, as shown in Figure 3A, namely PFOA and C/UOA based on the shape of the origami paper. The initial state of PFOA is inwardly curved, enabling expansion through applied pressure. A comparison between experimental and simulated trajectories with four different parameters is illustrated in Figure 3B. C/UOA is a type of internally wound actuator. If the thickness of a specific plane within the actuator is altered, its unfolding trajectories will be offset during the actuation process. The motion trajectories can be observed in Figure 3C.

FIG. 3.

Variation actuators of Z-shaped origami. (A) Design concepts and internal structural parameters of PFOA and C/UOA. (B) The motion trajectory of PFOA was compared with simulation data under different parameter settings (Groups 10a, 10b, 10c, and 10d) as the pressure varied. (C) The motion trajectory of C/UOA was compared with simulation data under different parameter settings (Groups 11a, 11b, 11c, and 11d) as the pressure varied. Note: In the PFOA experiment, the positive pressure P ranges from 0 to 100 kPa, and data points at 0, 40, and 100 kPa are sampled and compared. During the course of the PFOA experiment, positive pressure P varies from 0 to 60 kPa, with data points collected at 0, 20, and 60 kPa for comparison.

The simulation results are depicted in Figure S2 of Supplementary Data, and the different parameter configurations of PFOA and C/UOA can be referred to Tables S3 and S4 in Supplementary Data.

Structure design and kinematic analysis of TTOA

Inspiration from Kresling origami

Kresling origami is an origami with a torsional motion that occurs through a preset diagonal crease. Figure 4A shows our design idea of a TTOA. Most current studies based on Kresling origami are conventional with a square hexagonal cross section,^31,38–42 which can be twisted at an angle of 60° (360°/side length) in the ideal state (without considering the wall thickness). However, if a Kresling origami structure with a square cross section is designed, then the twistable angle in an ideal state is 90°, which can be twisted at a larger angle compared with the Kresling origami with a hexagonal cross section. Accordingly, a twisted tower origami was designed on the basis of the principle of traditional Kresling, and a TTOA was fabricated by 3D printing. The design idea of the TTOA can be viewed in Supplementary Movie S3.

FIG. 4.

Design of twist tower origami actuator. (A) The design concept and internal structural parameters of the TTOA. (B) Variation in contraction ratio and torsion angle of the actuator under different designated parameters: Groups 6a, 6b, and 6c involve changes in the side length W₂ of the actuator; Groups 6a, 7a, and 7b involve changes in the height h of the actuator cavity; Groups 6a, 8a, and 8 involve changes in the preset angle β; Groups 6a, 9a, and 9b involve changes in the wall thickness t₂ of the actuator. (C) Comparison between experimental data and simulation results for each set of actuators regarding contraction ratio and torsion angle. Note: During experimentation, vacuum pressure -P ranged from −80 to −0 kPa with sampling intervals at every −20 kPa.

Actuation mechanism of TTOA

The thickness of the top and bottom surfaces of the TTOA was greater than the thickness of four walls (during actuation, the top and bottom surfaces are stiffer than the four walls, making the actuator easier to compress), and a through-hole was designed at the bottom of the actuator to connect to the air tube, allowing the actuator to achieve a dual degree-of-freedom motion: contraction and torsion, driven by vacuum pressure −P. Unlike the vacuum-powered soft pneumatic actuator designed by Jin et al.,³² they achieved torsion by designing crease grooves in the four walls, which greatly increased the thickness of four walls of the actuator, correspondingly increasing the overall size of the actuator. However, the TTOA was designed with a preset tilt angle to achieve torsion, which not only reduced the 3D printing wall thickness material but also made the actuator highly compact. This actuator also exhibited bistability (movement from one stable state to another).

Parameters of TTOA

The square bottom side length of the TTOA, height, internal cavity height, four-wall thickness, and four-wall preset tilt angle were denoted as W₂, H, h, t₂, and β. When negative pressure actuation was performed, the height displacement of the actuator was ΔH, the contraction ratio was δ = ΔH/H, and the torsion angle was ϕ. Variation of the contraction ratio and torsion angle of the TTOA was investigated by setting the different parameters (Table S2 in Supplementary Data).

Kinematic analysis of TTOA

To investigate the kinematic characteristics of contraction and torsion in TTOA during driving, four actuator sets with different parameters were compared, and the results are presented in Figure 4B. Groups 6a, 6b, and 6c alter the length of the actuator's bottom edge. Under identical air pressure, an increase in the actuator's side length results in a higher contraction ratio. However, augmenting the bottom edge length (Groups 6a, 6b, and 6c) leads to a decrease in torsion angle. The transition between two stable states of this actuator occurs during “snap-through,” which refers to its shift from one stable state to another within an unstable structure at vacuum pressures mostly between −20 and −40 kPa (shaded gray region in Figure 4C).

The actuator's torque angle increment exhibited a small change with the increase in height for Groups 6a, 7a, and 7b, however, there was a significant reduction in contraction. The actuator's preset inclination angle is altered in Groups 6a, 8a, and 8b. Increasing the inclination angle enhances the initial state of the actuator's side wall fit, resulting in a narrower internal space, and this leads to a reduction in the folding rate of the wall, thereby decreasing both torsion angle and contraction magnitude. Interestingly, this adjustment promotes greater stability during the transition process between two steady states of the actuator while diminishing the prominence of “snap-through” phenomenon. Groups 6a, 9a, and 9b alter the thickness of the actuator's side wall, and an increase in wall thickness results in a delayed “snap-through” phenomenon, thereby reducing the response speed of the actuator. The comparison between the experimental and simulation results of the contraction ratio and torsion angle is illustrated in Figure 4C. The simulation details are provided in Figure S3 of Supplementary Data. The related experiments are presented in Supplementary Movie S4.

Analysis of Bionic Structure and Locomotion Mechanism of the Proposed Robot

Overall structure of the proposed robot

Figure 5A shows the overall structure of the herein designed robot with four origami actuators integrated into one robot body: (1) A C/UOA mounted on the head of robot; (2) the forelimb is assembled from the joint connector, TTOA, shoulder connector, and forelimb component; (3) the hind limb is modularly designed with ZSOA, foot component, and web component; and (4) the bottom of the robot is installed with a PFOA. Detailed information about the robot structure, Note S1 in Supplementary Data.

FIG. 5.

Design and analysis of frog-like leg actuator. (A) The robot uses a modular design in its overall structure. (B) The swimming process of frog can be divided into three stages: expansion, propulsion, and recovery. The articulations of both the frog and the robot (hip, knee, and ankle) have been delineated. A unidirectional water flow configuration has been implemented for the robot's webbing (highlighted by dot line). (C) (i) Comparison between changes in the angles of hip, knee, and ankle joints during a motion cycle for both the frog and the robot (expansion and propulsion stages). (ii) Change in angular velocity of the three robot joints. (D) Comparison of the simulated frog leg actuator's end motion trajectory during the expansion, propulsion, and recovery phases with the simulation results. Note: The experimental data on the angles of the frog's three joints were obtained through video recordings. The robot's imitation leg actuator operated under a driving condition with a maximum pressure of 120 kPa.

Design and analysis of the frog-like leg actuator

Swimming mechanism of frog

By observing freshwater frogs, we have identified their distinctive swimming patterns: expansion, propulsion, and recovery,^43,44 as shown in Figure 5B. During the expansion stage, the knee joint between the frog's thigh and calf is preferentially activated to adjust the position of limbs for a maximum range of webbed beating on water. In the propulsion stage, rapid activation of ankle and hip joints generates paddling force on water surface to promote frog swimming movement, utilizing inertia from this force, enabling gliding movement on water surface. Finally, during the recovery stage, all three joints of frog's legs are activated to return to the initial swimming position.

Parameter selection of ZSOA

Our objective is to develop a frog-like leg actuator by using ZSOA kinematic analysis and incorporating the mechanism of frog swimming locomotion. Consequently, we conducted tests to examine variations in joint angles during both the expansion and propulsion stages, defining the angles at the hip, knee, and ankle as θ_hip-F, θ_knee-F, and θ_ankle-F. As shown in Figure 5C, the knee joint exhibits the most rapid increase in angle throughout the entire motion cycle, indicating its early activation. Moreover, θ_knee-F eventually approaches 140°, while the ankle angle is slightly smaller than that of knee, with θ_ankle-F tending toward 50°. The variation in joint angle of the frog during locomotion is illustrated in Figure S5 of Supplementary Data. Based on these experimental findings, we select ZSOA parameters.

The kinematic analysis reveals that the length of actuator channel is directly proportional to the expansion speed of flexible cavity. Therefore, the parameter c₂ of ZSOA is set to 12 mm, and this channel is used as the knee joint of frog. Similarly, c₃ is set to 3 mm as the ankle joint, with the intention of ensuring a lower angular response speed compared with that of the knee joint. As for c₁, it is set at 6 mm to represent the hip joint, this choice is motivated by observing that the hip joint angle response lags behind the knee joint angle response and ultimately reaches a development angle of 50°. The actuator length L is set to 80 mm and the tilt angle α is set to 90°, as increasing the range of motion will enhance the efficiency of water surface paddling.

In addition, considering both response speed and useful life, a wall thickness t₁ of 1.2 mm is chosen for the actuator. The dimensional parameters of frog-like leg actuator are shown in Figure S4A of Supplementary Data.

Aquatic locomotion mechanism of the robot

The frog-like leg actuator is carefully chosen based on specific parameters, and an in-depth investigation is the leg driving trajectory of the aquatic locomotion mechanism used by the robot. The angular positions of the robot's hip, knee, and ankle joints were denoted as θ_hip-R, θ_knee-R, and θ_ankle-R, respectively. Subsequently, the changes in joint angles for each component of the frog leg actuator were recorded under actuation pressure of 120 kPa. The initial state of the joint angle during the movement cycle is inconsistent, as depicted in Figure 5C, due to variations in the installation position of the imitation leg actuator on our designed robot compared with that of a frog's thigh. However, there exists a highly similar changing trend between the robot's joint angle and the frog's swimming motion, the maximum achievable values for all three joint angles closely resemble those observed in real frogs.

Furthermore, we conduct angular velocity tests on each joint of the robot and observed that at 1.75 s, the knee joint exhibited a peak angular velocity, surpassing both the ankle and hip joints. This finding provides further evidence supporting the activation sequence of the robot's knee joint aligning with the swimming characteristics observed in frogs.

The end trajectory of the frog leg actuator under driving conditions is depicted in Figure 5D. During the expansion stage (green line), a noticeable increase in x-axis trajectory indicates activation of the knee joint, leading to a rapid rise in θ_knee-R. Upon entering the propulsion stage (orange line), there is a sharp elevation in y-axis trajectory, signifying activation of both ankle and hip joints. Finally, during the recovery stage (blue line), the trajectory differs from that observed during propulsion, possibly due to the unique elastic resilience properties of the material. Moreover, the robot's webbed structure is designed to function unidirectionally, and during the propulsion stage, the webbed area maximizes contact with water, while in the recovery phase, it gets effortlessly washed away by the current. This design minimizes resistance during swimming and prevents any backward movement. Robot imitates frog swimming in Supplementary Movie S5.

Design and analysis of mudskipper-like joint and pelvic fin actuators

Crawling mechanism of mudskipper

The mudskipper, despite being an aquatic creature, possesses the capability to traverse terrestrial environments. The process of land crawling in mudskippers encompasses the following two distinct stages: propulsion and recovery. As shown in Figure 6A, in the propulsion stage, mudskippers rely on robust joints to torsion their pectoral fins, generating ground thrust and facilitating locomotion. In addition, the elongated pelvic fins located at the ventral side of their body provide auxiliary support to the pectoral fins, as some scholars propose that will enhance the generated thrust.^45,46 During the recovery phase, the mudskipper elevates its pectoral fin back to the initial state without altering its body position.

FIG. 6.

Design and analysis of mudskipper-like joint and pelvic fin actuators. (A) Inspired from the locomotion mechanism of mudskippers, forward and backward locomotion modes have been devised for the robot. (B) Comparison of joint angle variations between mudskipper and robot during the propulsion stage. (C) Comparison of joint angle variations between mudskipper and robot during the recovery stage. (D) Comparative analysis of the single-cycle displacement of the robot with different pelvic fin actuator parameters (Groups 10a–10d). (E) Comparison of single-cycle turning angles of the robot under different pelvic fin actuator parameters (Groups10a–10d). Note: Experimental data on the joint angles of mudskippers during the propulsion and recovery were acquired through video analysis. The pelvic fin actuator of the robot operates under actuation pressures of p = 0, 50, and 100 kPa, while the joint twisting actuators operate under an actuation pressure of −80 kPa.

Parameter selection of TTOA

One of this study's objectives is to develop an actuator that mimics the joint structure and crawling motion mechanism of mudskippers, utilizing the TTOA kinematic analysis method. Therefore, we conduct experimental analysis to investigate the torsion angle changes of mudskipper joints during both propulsion and recovery stages, denoted as θ_joint-M (as illustrated in Fig. 6B, C). The obtained measurement data revealed that the maximum torsion angle of mudskipper joint is ∼70°. The snap-shots of mudskipper's joint angle changes are depicted in Figure S6 of Supplementary Data. Based on these data, we chose the parameters for TTOA. According to the maximum torsion angle of the mudskipper joint, the cavity height (h) of TTOA is set at 30 mm, and the preset tilt angle (β) is set to 85°.

To account for actuator volume and response speed, the bottom side length (W₂) and wall thickness (t₂) of the actuator are adjusted to 25 and 0.8 mm, respectively. The dimensional parameters of mudskipper-like joint actuator are shown in Figure S4B of Supplementary Data.

The torsion angle of the robot joint actuator is denoted as θ_joint-R. Comparing it with the joint torsion angle of the mudskipper during propulsion and recovery, a highly similar trend between them as depicted in Figure 6B. The advancement of θ_joint-R over θ_joint-M can be attributed to the occurrence of the “snap-through” phenomenon induced by actuator actuating. However, in Figure 6C, the mudskipper in the recovery stage exhibits a rapid pectoral fin, leading to an in θ_joint-M per unit time and causing it to significantly lag behind θ_joint-R.

In summary, both movements tend to converge toward the same final limit value for joint angle, thereby inspiring the forward locomotion of our robot. In Figure 6C and D, the single-cycle displacement and turning angle of the robot were tested under different parameters (Groups 10a–10d) for the pelvic fin actuator. The Group 10a pelvic fin actuator demonstrates excellent backward ability when unactuated. When fully actuated (p = 100 kPa), the forward displacement caused by the Group 10a actuator is slightly lower than that of Group 10b. Consequently, augmenting the channel length (c₄) of the pelvic fin actuator can enhance the robot's forward propulsion, however, significantly impeding its backward locomotion. In corner tests at an actuation pressure of 100 kPa, three actuators (Groups 10a, 10b, and 10c) exhibit good turning performance.

Consequently, considering combined capabilities in terms of forward motion, backward motion, and turning ability led to the choosing of pelvic fin actuator variant represented by Group 10a.

Parameter selection of PFOA

We aim to develop a pelvic fin actuator capable of enhancing the thrust generated by the joint actuator toward the ground, thereby augmenting the robot's terrestrial locomotion capabilities. In Figure 6D and E, pelvic fin actuators in Groups 10a–10b are compared. Considering the factors of the robot's forward, backward, and turning locomotion, the actuator Group 10a is selected. The dimensions for top surface (L₁), bottom surface (L₂), and channel (c₄) are 29.7, 29.5, and 2.5 mm, respectively. The dimensional parameters of mudskipper-like pelvic fin actuator are shown in Figure S4C of Supplementary Data.

Terrestrial locomotion mechanism of the robot

The principle of the robot's terrestrial locomotion is devised by carefully selecting parameters for both the joint and pelvic fin actuators, as well as their coordinated actuation (Fig. 6A). The robot exhibits the following three modes of terrestrial locomotion: forward, turning, and backward. (1) During forward locomotion, the pelvic fin actuator remains actuated throughout, causing elevation of the lower end of the robot body. In the propulsion stage, two joint actuators are rotated to drive the pectoral fin actuators at their extremity, generating propulsive force on the ground and facilitating displacement of the robot by sliding the pelvic fin actuator. During the recovery stage, the joint actuators return to the initial state, while the pelvic fin actuator prevents the robot from moving backward due to its anisotropic friction.

(2) In turning locomotion, the underlying principle resembles forward motion and can be propelled by a solitary joint actuator. (3) In backward locomotion, the pelvic fin actuator remains unactuated throughout the entire motion process. During the propulsion stage of backward locomotion, joint actuator drives the pectoral fin actuator to generate ground resistance, resulting in an elevated robot body position without actual movement. Subsequently, in the recovery stage, backward locomotion is initiated by utilizing propulsion force generated through pectoral fin actuator recovery on the ground. The robot imitates the mudskipper crawling in Supplementary Movie S6.

Design and analysis of frog-like tongue actuator

Hunting mechanism of frog

During hunting, frogs typically exhibit two distinct states as follows: tongue projection and retraction.⁴⁷ As depicted in Figure 7A, the rapid rotation of the tongue is facilitated by robust muscles in the lower jaw during projection, enabling it to adhere mucus upon contact with the insect. When retracted, the tongue retrieves the insect to the mouth, restoring to its initial state.

FIG. 7.

Design and analysis of frog-like tongue actuator. (A) Motion analysis of the frog's tongue hunting mechanism and the robot's tongue hunting. (B) Velocity and acceleration data of the frog tongue's tip obtained from video during both the projection and retraction stages. (C) Comparison of the velocity of the designated points at the frog-like tongue actuator's end in Groups 11a–11d with respect to time variation. (D) Comparison of the accelerations of the designated points at the frog-like tongue actuator's end in Groups 11a–11d with respect to time variation. Note: An actuation pressure of 100 kPa is applied for experiments on actuator velocity and acceleration.

Parameter selection of C/UOA

Based on the analysis of frog hunting data from the video, we observe that the velocity during projection is significantly higher than retraction. Furthermore, the acceleration variation of tongue during projection is more obvious than retraction, as illustrated in Figure 7B. The trajectory of the frog's tongue during predation is depicted in Figure S7 of Supplementary Data. To develop an actuator that mimics the frog tongue projection mechanism, the C/UOAs of Groups 11a–11d are used to evaluate its velocity and acceleration characteristics. Subsequently, the Group 11a actuator is chosen, with parameters t₃, t₄, and t₅ set at 0.8 mm. The dimensional parameters of frog-like tongue actuator are shown in Figure S4D of Supplementary Data.

Hunting mechanism of the robot

Lin et al. designed a dielectric elastomeric fluid sensor-driven bionic soft proboscis actuator, for coiling and uncoiling motions at an external voltage, but it is complex to fabricate and requires an external high-voltage power supply.⁴⁸ The frog-like tongue actuator is positioned on the robot's head (Fig. 7A). During the projection stage, the actuator rapidly uncoils and adheres to the object using an adhesive tape at its tip. In the retraction stage, the elastic resilience of the material allows the actuator to retract and bring back the object into its body. In Figure 7C, D, we conduct tests on the velocity and acceleration of the designated point at the extremity of the actuator. Results show that the projection mechanism is completed by the Group 11a actuator within a remarkable duration of 0.25 s, exhibiting a significantly superior speed compared with actuators with other parameters.

Furthermore, its acceleration reached a maximum value of 12 m/s⁻², which closely aligns with the rapid projection mechanism observed in frog tongues and consequently influenced our selection of the 11a actuator. The robot imitating frog hunting is presented in Supplementary Movie S7.

Experiment on Robot Locomotion Capability

Hardware control system and gait of robot

Six distinct external air circuits are necessary for the robot's locomotion, which are regulated by six solenoid valves. In Figure 8A, Valves 4 and 5 correspond to joint actuators and receive negative pressure from a vacuum pump. Valves 1 and 2 correspond to leg actuators, Valve 3 corresponds to the pelvic fin actuator, and Valve 6 corresponds to the tongue actuator, which is supplied with positive pressure from a system with activation and deactivation of relay power to regulate the opening and closing of the air pathway. The hardware setup used for control is illustrated in Figure S8 of Supplementary Data.

FIG. 8.

Hardware control system and gait of robot. (A) Connection method of the hardware control system. (B) Nine gait designs of the robot. Note: Dark gray indicates a positive pressure pump that provides positive pressure to the valve, medium gray denotes a vacuum pump that provides negative pressure to the valve, and light gray means no gas supply. The bidirectional arrows indicate the transition between the two modes in one cycle.

The gaits 1 (forward swimming), 4 (forward crawling), and 9 (tongue hunting) are derived from the biological locomotion postures, as depicted in Figure 5B. Building upon the robot's inherent structural design and extensive locomotion experimentation, additional six gait modes have been developed. Consequently, a total of nine distinct gait patterns are showcased in Supplementary Movie S8.

Robotic swimming proficiency assessment

Swimming forward velocity test

The duration of one complete cycle of the robot from the initiation of movement to returning to its initial state is denoted as T (i.e., cycle T = 1/f), while the internal positive pressure within the actuator is represented by P. To determine the maximum swimming speed of the robot, experiments were conducted by varying both inflation pressures (P) and frequencies (f) and analyzing their effects on swimming velocity (Fig. 9C, D). The results demonstrate that at an inflation pressure of p = 120 kPa and a test duration of t = 15 s, the robot achieves a maximum swimming distance of 1200 mm with a corresponding swimming velocity of 80 mm/s (i.e., equivalent to 0.6 BL/s).

FIG. 9.

Robotic swimming proficiency assessment. (A) A demonstration of a robot forward through aquatic locomotion. (B) A demonstration of a robot turning through aquatic locomotion. (C) The forward swimming velocity of the robot was compared under a different pressure P, with a specified frequency f = 0.5 Hz and a duration of t = 15 s in the experiment. (D) The forward swimming velocity of the robot was compared at different frequencies f, with a fixed pressure p = 120 kPa and a duration of t = 15 s in the experiment. (E) The comparison of robot turning angles under different frequencies f reveals that the experiment specifies a pressure P of 120 kPa and a time t of 5 s. Note: The dot on the robot is the trace mark point.

Swimming turning velocity test

The robot's aquatic maneuverability was assessed to demonstrate its flexible swimming capabilities in water. The impact of different frequencies on the robot's turning was compared under a pressure condition of 120 kPa (Fig. 9E). The results indicate that at a frequency of 0.5 Hz, the maximum turning angle reached 53.7° within a time t = 5 s. Further analysis of the swimming experiments can be found in Note S2 of Supplementary Data, while Supplementary Movie S9 showcases additional robotic swimming.

Locomotion energy efficiency analysis in swimming

The energy efficiency of mobile soft robots was assessed by Shui et al.,⁴⁹ taking account the efficiency of energy, actuators, and locomotion. The effective motion energy of a soft robot is defined as follows: $E_{l o c o} = \frac{1}{2} M \cdot {\bar{v}}^{2}$ (1)

Where M represents the total mass of the robot and $\bar{v}$ denotes the average translational velocity of the mass center in the desired locomotion direction. According to the prescribed formula, the robot's maximum achievable locomotion energy during swimming, $E_{l o c o} = \frac{1}{2} \times 208.3 g \times {(0.08 m s^{- 1})}^{2} = 666.56 μ J$ . The output power of the positive pressure pump can be determined using the following formula, considering that the robot operates in a pneumatic control mode: $E_{1} = P \cdot Q$ (2)

The output power of the positive pressure pump can be calculated by multiplying the supply pressure (P) and the flow rate of the air pump (Q), yielding $E_{1} = 120 k P a \times 20 L m i n^{- 1} = 40 W$ . Considering the robot's swimming period (T = 2 s), the total work output of the positive pressure pump in a single cycle can be determined as $E_{m} = E_{1} \cdot T = 40 W \times 2 s = 80 J$ . The locomotion energy efficiency of a soft mobile robot is defined as follows: $η_{l o c o} = \frac{E_{l o c o}}{E_{m}} \cdot 100 %$ (3)

Therefore, the maximum efficiency of locomotion energy for the swimming robot can be calculated as $η_{l o c o} = 666.56 μ J ∕ 80 J \times 100 % = 0.000833 %$ .

Robotic crawling and hunting proficiency assessment

Crawling forward velocity test

To evaluate the terrestrial crawling velocity of the robot, various frequencies (f) are used to investigate their impact on crawling velocity in gait 4 (Fig. 10A). The results indicate that at f = 1.67 Hz, gait 4 exhibits a higher crawling velocity compared with other frequencies. In addition, an alternative gait, namely gait 5, has been developed and previous studies have highlighted its advantages.^50,51 With the implementation of gait 5 crawling style, the robot achieves a maximum velocity of 20.5 mm/s within a time frame of t = 10 s.

FIG. 10.

Robotic crawling and hunting proficiency assessment. (A) Comparison of robot crawling distance was conducted at various frequencies (f) with a fixed test time (t = 10 s) during the experiment. (B) Comparison of robot crawling and turning angles at different frequencies (f) during a 5-s test period in the experiment. (C) The robot traces the trajectory of a circular path with a diameter measuring 70 cm. (D) The experiment involved comparing the crawling distance of the robot under different loads, with a test frequency of f = 1.25 Hz and a duration of t = 10 s. (E) The robot's crawling distance in various terrains was compared. During the experiment, a test frequency of f = 1.67 Hz and a duration of t = 10 s were utilized. (F) The robot is equipped with a camera mounted on its head and transports the captured insects to their designated destination. Note: All crawling tests are conducted under negative pressure of −80 kPa and positive pressure of 100 kPa. The dot represents the designated track marking point.

Crawling turning velocity test

The robot's crawling turning angle (Fig. 10B) was examined to investigate the impact on turning velocity at different frequencies f, with a fixed time of t = 5 s. The results revealed that higher frequencies enhanced the crawling velocity of the robot, enabling to achieve a velocity of 30° s⁻¹ at a frequency of 2 Hz.

Evaluation of locomotion capabilities

In addition, to demonstrate precise control capabilities, remote operation was used to guide the robot along a path with a diameter of 70 cm, as depicted in Figure 10C. The carrying capacity of the robot was examined in Figure 10D, and a comparison was made between the crawling velocities of the robot, while carrying different weights at a frequency of f = 1.25 Hz, the experiment demonstrates that the robot's crawling velocity is the fastest when carrying a 100 g load, compared with other conditions within the specified 10-s time frame. The weight of a single TTOA is merely 7 g, yet it exhibits an impressive load-bearing capacity of 200 g, showcasing the remarkable structural robustness of the Kresling origami design. In Figure 10E, the robot traversed diverse terrains composed of different materials at a frequency f = 1.67 Hz, demonstrating its capability to navigate through complex environments.

Supplementary Movie S10 provides visual documentation of these experiments. Figure 10F illustrates the controlled execution of an insect-catching task by the robot, which successfully transported the captured insect to its intended destination. Further analysis and discussion on both crawling and hunting experiments can be found in Note S3 of Supplementary Data. Supplementary Movie S11 presents a detailed depiction of the robot's performance in capturing insects.

Locomotion energy efficiency analysis in crawling

The Equations (1), (2), and (3) can be utilized to compute the maximum locomotion energy of crawling robot, denoted as $E_{l o c o} = \frac{1}{2} \times 208.3 g \times {(0.0206 m s^{- 1})}^{2} = 44.2 μ J$ . In addition, the output power of both the positive pressure and negative pressure pumps is given by $E_{1} = 100 k P a \times 20 L m i n^{- 1} + 80 k P a \times 75 L m i n^{- 1} = 133.3 W$ . While the total output work in a single cycle is represented as $E_{m} = E_{1} \cdot T = 133.3 W \times 1.2 s = 160 J$ . Consequently, the efficiency of maximum locomotion energy during crawling can be expressed as $η_{l o c o} = 44.2 μ J ∕ 160 J \times 100 % = 0.0000276 %$ . Compared with the soft quadruped robot proposed by Tolley et al.⁵¹ and Shepherd et al.,⁵² the energy efficiency of crawling locomotion is comparable. The energy efficiency of swimming is significantly higher than that of crawling, as evidenced by the findings.

Robotic amphibious proficiency assessment

Amphibious robots possess the ability to seamlessly transition between land and water, thereby demonstrating their remarkable amphibious switching capability (Fig. 11A, B). The key challenge lies in minimizing the time required for mode switching during amphibious motion. As the robot transitions from land to water, it smoothly switches from gait 5 to gait 1 when its head floats on the water surface. Conversely, when transitioning from water to land, once its head makes contact with solid ground and can no longer swim, it effortlessly shifts from gaits 1 to 5. Supplementary Movie S12 illustrates the experimental demonstration of amphibious transition by the robot. In addition, Figure 11C showcases a simulated experiment depicting the robot's successful transition in a complex outdoor terrain (Supplementary Movie S13). Further details are provided in Note S4 of Supplementary Data.

FIG. 11.

Composite images featuring sequential snapshots showcasing the amphibious locomotion capability of the soft robot are presented. (A) The robot's land to water is demonstrated in an amphibious manner. (B) The robot successfully lands on land after being in water, exhibiting its amphibious capabilities. (C) An outdoor experiment conducted at NUIST Jinniu Lake Campus Pond illustrates the robot's seamless transition from land to water.

Discussion

Experimental validation demonstrates the viability and practicality of a bioinspired origami amphibious soft robot, which is based on the fabrication of origami actuators. The subsequent discussion encompasses the performance, characteristics, and functionality of this robotic system in terms of its structural composition, material properties, and operational capabilities.

Structural composition

The origami technology is a highly advantageous tool for transforming 2D structures into three dimensions, and the topological combinations of origami have demonstrated significant benefits such as large deformation,⁵³ high load-bearing capacity,⁵⁴ programmability,⁵⁵ and lightweight characteristics.⁵⁶ Four origami actuators are designed based on the Miura and Kresling origami prototypes: ZSOA, PFOA, and C/UOA. In addition, a novel concept that combines origami with bionics is proposed, in which these actuators mimic the postures of living creatures. This study confirms the substantial advantages of utilizing origami structures in a bioinspired design. Furthermore, it provides valuable insights and references for designing other bioinspired structures.

Material properties

The robot's origami actuator is constructed entirely from the flexible material TPE, while the remaining components are composed of polylactic acid. Using a 3D-printed modular design, this robot effectively mitigates manufacturing and assembling complexities. Given that most origami structures incorporate hollow cavities, they exhibit exceptional material efficiency. Remarkably, the combined weight of all origami actuators amounts to merely 90 g, which equates to <20 Renminbi.

Operational capabilities

The experiments have demonstrated that the robot designed not only had the ability to navigate forward and execute left and right on water, but also had the capability to move forward, right, as well as backward on land. Moreover, this robot had exceptional amphibious locomotion across different mediums: transitioning seamlessly from land to water and vice versa. In addition, it is capable of carrying heavy utilizing its tongue for insect hunting, avoiding obstacles effectively, and crawling across various terrains. To our knowledge, among the current bioinspired soft robots with amphibious capabilities, this the most comprehensive integration of both aquatic and terrestrial functionalities (Table 1 provides a comparative analysis of different bioinspired soft amphibious robots; “−” indicates unreported performance details in the article). The performance comparison with other robots can be found in Note S5 of Supplementary Data.

Table 1.

Comparison of Different Bionic Soft Amphibious Robots

Robot	Bionic posture	Actuation method	Different actuators for water and land locomotion	Different actuation modes for water and land locomotion	Water speed (BL/s)	Turning speed on water (° s⁻¹)	Land speed (BL/s)	Turning speed on land (° s⁻¹)	Function
Tang et al.¹¹	Geometrid wriggling	Pneumatic (positive pressure)	No	No	—	—	0.03	—	Climbing in water and on land
Paschal et al.¹²	Sea urchin crawling	Pneumatic (positive pressure) + magnetic adsorption	No	No	—	—	0.03	25	Adsorption on the iron surface forward and turning
Li et al.¹³	Dog walking and paddling	Pneumatic (positive pressure, precharged) + motor	No	No	0.05	—	0.49	—	Advance in water and on land
Wu et al.¹⁵	Turtle crawling	Pneumatic (positive + negative pressure)	No	Yes	0.1	—	0.97	25.4	Water advance, land advance, turn, and amphibious landing
Faudzi et al.¹⁶	Lizard crawling	Pneumatic (positive pressure)	No	No	0.1	—	0.27	—	Advance in water and on land
Milana et al.¹⁷	Geometrid wriggling and eel swimming	Pneumatic (positive pressure)	No	Yes	0.05	—	0.03	—	Advance in water and on land
Dong et al.¹⁸	Geometrid wriggling and human swimming	Pneumatic (positive + negative pressure)	Yes	Yes	0.7	—	0.003	—	Advance in water and on land and perceive obstacles
Hwang et al.¹⁹	Seals crawling and swimming	SMA	Yes	Yes	0.07	6.07	0.009	0.46	Forward and turn in water, forward and turn on land, and transition from land to water
Baines et al.²⁰	Sea turtles swimming and land turtle walking	Pneumatic (positive pressure) + motor + heating of thermosetting materials	No	Yes	0.26	—	0.09	—	In-water and land advancement and transition from land to water
Xia et al.⁵⁷	Fish and snake swimming	Motor	No	No	0.36	—	0.09	18	Underwater forward, land forward, backward, turn, and amphibious landing
Patel et al.⁵⁸	Quadruped crawling and frog swimming	SMA	No	Yes	0.06	—	0.07	—	Forward and flipping in water and forward on land
Our robot	Frog swimming, mudskipper crawling, and frog hunting	Pneumatic (positive + negative pressure)	Yes	Yes	0.6	10.74	0.16	30	Forward and turning on water, forward, turning, and backward on land, carrying heavy objects, hunting insects, avoiding obstacles, switching back and forth between land and water transitions, and crawling on different terrains

BL/s, body length per second; SMA, shape memory alloy.

Conclusion and Future Work

This work presents the design and fabrication of a 3D-printed origami actuator for a bionic soft robot with amphibious locomotion and tongue hunting capabilities. The strengths of this work include the following: first, the development of four unique origami actuators based on the prototypes of Miura and Kresling origamis, including a ZSOA that mimics a frog swimming in water, TTOA, and PFOA, simulating the land crawl of a mudskipper, and an actuator (C/UOA) for the coiling/uncoiling movement of a hunting tongue. Second, a comparative parametric analysis of the motion posture of the origami actuator was performed to optimize robot performance. Third, a soft robot prototype was fabricated, and nine robot gaits were designed and tested through a series of experiments to demonstrate its powerful yet integrated amphibious capabilities and potential functions, such as hunting.

This work contributes to the development of advanced soft robotics with potential applications in environmental monitoring, search and rescue operations, and so on. In future research, the autonomous control of the robot's locomotion and amphibious transition will be investigated based on sensory feedback. An untethered robotic system design would also broaden the robot's application scenarios. Furthermore, sensors testing water quality such as pH value and turbidity would be integrated into the amphibious robot so that it could perform water quality monitoring tasks in the land–water transition zone.

Footnotes

Author Disclosure Statement

No interests to disclose.

Funding Information

This research was supported, in part, by the National Natural Science Foundation of China (Grant No. 52005269), the Research Project of State Key Laboratory of Mechanical System and Vibration MSV202319, and the GuangDong Basic and Applied Basic Research Foundation No. 2022A1515110067.

Supplementary Material

References

Ren

, Yu

. Research status of bionic amphibious robots: A review. Ocean Eng, 2021; 227:108862.

Hamidi

, Almubarak

, Rupawat

, et al. Poly-Saora robotic jellyfish: Swimming underwater by twisted and coiled polymer actuators. Smart Mater Struct, 2020; 29(4):045039.

, Chen

, Zhou

, et al. Self-powered soft robot in the Mariana Trench. Nature, 2021; 591(7848):66–71.

Fan

, Wang

, Yu

, et al. Swimming performance of the frog-inspired soft robot. Soft Robot, 2020; 7(5):615–626.

, Zheng

, Liu

, et al. Glowing Sucker Octopus (Stauroteuthis syrtensis)—Inspired soft robotic gripper for underwater self-adaptive grasping and sensing. Adv Sci, 2022; 9(17):2104482.

Kandhari

, Wang

, Chiel

, et al. An analysis of peristaltic locomotion for maximizing velocity or minimizing cost of transport of earthworm-like robots. Soft Robot, 2021; 8(4):485–505.

Zhang

, Liu

, Zhou

, et al. Crab-inspired compliant leg design method for adaptive locomotion of a multi-legged robot. Bioinspir Biomim, 2022; 17(2):025001.

, Shi

, Pinto

, et al. A novel pneumatic soft snake robot using traveling-wave locomotion in constrained environments. IEEE Robot Autom Lett, 2020; 5(2):1610–1617.

Natarajan

, Chia

, Faudzi

AAM

, et al. Bio inspired salamander robot with Pneu-Net Soft actuators—Design and walking gait analysis. Bull Polish Acad Sci Techn Sci, 2021; 69:e137055.

10.

Baines

, Fish

, Kramer-Bottiglio

. Amphibious robotic propulsive mechanisms: Current technologies and open challenges. In: Bioinspired Sensing, Actuation, and Control in Underwater Soft Robotic Systems. ( Paley

, Wereley

NM.

eds.) Springer: Cham; 2021; pp. 41–69.

11.

Tang

, Zhang

, Lin

, et al. Switchable adhesion actuator for amphibious climbing soft robot. Soft Robot, 2018; 5(5):592–600.

12.

Paschal

, Bell

, Sperry

, et al. Design, fabrication, and characterization of an untethered amphibious sea urchin-inspired robot. IEEE Robot Autom Lett, 2019; 4(4):3348–3354.

13.

, Fish

, Chen

, et al. Bio-inspired robotic dog paddling: kinematic and hydro-dynamic analysis. Bioinspir Biomim, 2019; 14(6):066008.

14.

Bell

, Weaver

, Wood

. An ambidextrous starfish-inspired exploration and reconnaissance robot (the ASTER-bot). Soft Robot, 2022; 9(5):991–1000.

15.

, Xu

, Zhao

, et al. A fully 3D-printed tortoise-inspired soft robot with terrains-adaptive and amphibious landing capabilities. Adv Mater Technol, 2022; 7(12):2200536.

16.

Faudzi

AAM

, Razif

MRM

, Endo

, et al. Soft-amphibious robot using thin and soft McKibben actuator. In: 2017 IEEE International Conference on Advanced Intelligent Mechatronics (AIM), Munich, Germany; IEEE; 2017; pp. 981–986; doi: 10.1109/AIM.2017.8014146.

17.

Milana

, Van Raemdonck

, Cornelis

, et al. EELWORM: A bioinspired multimodal amphibious soft robot. In 2020 3rd IEEE International Conference on Soft Robotics (RoboSoft), New Haven, CT, USA; 2020; pp. 766–771.

18.

Dong

, Yang

, Ding

, et al. Bioinspired amphibious origami robot with body sensing for multimodal locomotion. Soft Robot, 2022; 9(6):1198–1209.

19.

Hwang

, Wang

. Shape memory alloy-based soft amphibious robot capable of seal-inspired locomotion. Adv Mater Technol, 2022; 7(6):2101153.

20.

Baines

, Patiballa

, Booth

, et al. Multi-environment robotic transitions through adaptive morphogenesis. Nature, 2022; 610(7931):283–289.

21.

Son

, Park

, Na

, et al. 4D multiscale origami soft robots: A review. Polymers, 2022; 14(19):4235.

22.

Kim

, Byun

, Kim

J-K

, et al. Bioinspired dual-morphing stretchable origami. Sci Robot, 2019; 4(36):eaay3493.

23.

Kim

, Eom

, Cho

K-J

. A dual-origami design that enables the quasisequential deployment and bending motion of soft robots and grippers. Adv Intell Syst, 2022; 4(3):2100176.

24.

Dorsey

, Huang

, Wen

. Origami-patterned capacitor with programmed strain sensitivity. Multifunct Mater, 2022; 5(2):025001.

25.

, Chang

, Narumi

, et al. Printable Origami Bistable Structures for Foldable Jumpers. In 2022 International Conference on Robotics and Automation (ICRA), Philadelphia, PA, USA; 2022; pp. 7131–7137.

26.

Liang

, Gao

, Huang

, et al. Design of a Rigid-Flexible Coupling Origami Gripper. In 2021 IEEE International Conference on Robotics and Biomimetics (ROBIO), Sanya, China; 2021; pp. 1283–1287.

27.

, Stampfli

, Xu

, et al. A vacuum-driven origami “magic-ball” soft gripper. In 2019 International Conference on Robotics and Automation (ICRA), Montreal, QC, Canada; 2019; pp. 7401–7408.

28.

Kim

, Seo

, Yu

, et al. Deployable Soft Pneumatic Networks (D-PneuNets) Actuator with Dual-Morphing Origami Chambers for High-Compactness. IEEE Robot Autom Lett, 2021; 7(2):1262–1269.

29.

Hilby

, Padia

, Hunter

Design and Analysis of Origami-Inspired, Large-Elongation, Reconfigurable Soft Robot Modules. In 2022 IEEE 5th International Conference on Soft Robotics (RoboSoft), Edinburgh, United Kingdom; 2022; pp. 132–139.

30.

Zhang

, Fang

, Xu

. Yoshimura-origami based earthworm-like robot with 3-dimensional locomotion capability. Front Robot AI, 2021; 8:738214.

31.

, Wu

, Nishikawa

, et al. Soft robotic origami crawler. Sci Adv, 2022; 8(13):eabm7834.

32.

Jin

, Li

, Wang

, et al. Origami-inspired soft actuators for stimulus perception and crawling robot applications. IEEE Trans Robot, 2021; 38(2):748–764.

33.

Melancon

, Forte

, Kamp

, et al. Inflatable origami: Multimodal deformation via multistability. Adv Funct Mater, 2022; 32(35):2201891.

34.

Miyashita

, Guitron

, Ludersdorfer

, et al. An untethered miniature origami robot that self-folds, walks, swims, and degrades. In 2015 IEEE international conference on robotics and automation (ICRA), Seattle, WA, USA; 2015; pp. 1490–1496.

35.

Sayed

, Roberts

, McKenzie

, et al. Limpet II: A modular, untethered soft robot. Soft Robot, 2021; 8(3):319–339.

36.

Suzuki

, Zheng

, Kakehi

, et al. Shapebots: Shape-changing swarm robots. In Proceedings of the 32nd annual ACM symposium on user interface software and technology, New Orleans, LA, USA; 2019; pp. 493–505.

37.

Tachi

, Miura

. Rigid-foldable cylinders and cells. J Int Assoc Shell Spatial Struct, 2012; 53(4):217–226.

38.

, Fan

, Zhu

, et al. Origami-inspired soft twisting actuator. Soft Robot, 2022; 10(2):395–409.

39.

, Wu

, Dai

, et al. Spinning-enabled wireless amphibious origami millirobot. Nat Commun, 2022; 13(1):3118.

40.

Bhovad

, Kaufmann

, Li

. Peristaltic locomotion without digital controllers: Exploiting multi-stability in origami to coordinate robotic motion. Extreme Mech Lett, 2019; 32:100552.

41.

, Kidambi

, Wang

, et al. Uncovering rotational multifunctionalities of coupled Kresling modular structures. Extreme Mech Lett, 2020; 39:100795.

42.

, Ze

, Dai

, et al. Stretchable origami robotic arm with omnidirectional bending and twisting. Proc Natl Acad Sci, 2021; 118(36):e2110023118.

43.

Nauwelaerts

, Aerts

. Propulsive impulse as a covarying performance measure in the comparison of the kinematics of swimming and jumping in frogs. J Exp Biol, 2003; 206(Pt 23):4341–4351.

44.

Aerts

, Nauwelaerts

. Environmentally induced mechanical feedback in locomotion: Frog performance as a model. J Theoret Biol, 2009; 261(3):372–378.

45.

Pace

, Gibb

. Mudskipper pectoral fin kinematics in aquatic and terrestrial environments. J Exp Biol, 2009; 212(14):2279–2286.

46.

Swanson

, Gibb

. Kinematics of aquatic and terrestrial escape responses in mudskippers. J Exp Biol, 2004; 207(Pt 23):4037–4044.

47.

Kleinteich

, Gorb

. Frog tongue acts as muscle-powered adhesive tape. R Soc Open Sci, 2015; 2(9):150333.

48.

Lin

P-W

, Liu

C-H

. Bio-inspired soft proboscis actuator driven by dielectric elastomer fluid transducers. Polymers, 2019; 11(1):142.

49.

Shui

, Zhu

, Yang

, et al. Energy efficiency of mobile soft robots. Soft Matter, 2017; 13(44):8223–8233.

50.

Drotman

, Jadhav

, Sharp

, et al. Electronics-free pneumatic circuits for controlling soft-legged robots. Sci Robot, 2021; 6(51):eaay2627.

51.

Tolley

, Shepherd

, Mosadegh

, et al. A resilient, untethered soft robot. Soft Robot, 2014; 1(3):213–223.

52.

Shepherd

, Ilievski

, Choi

, et al. Multigait soft robot. Proc Natl Acad Sci USA, 2011; 108(51):20400–20403.

53.

Park

, Kim

, Yu

S-Y

, et al. Deployable soft origami modular robotic arm with variable stiffness using facet buckling. IEEE Robot Autom Lett, 2022; 8(2):864–871.

54.

Zhu

, Fan

, Wu

, et al. Soft robots for cluttered environments based on origami anisotropic stiffness structure (OASS) inspired by desert iguana. Adv Intell Syst, 2023; 5(6):2200301.

55.

Jin

, Forte

, Deng

, et al. Kirigami-inspired inflatables with programmable shapes. Adv Mater, 2020; 32(33):2001863.

56.

Kim

S-R

, Lee

D-Y

, Ahn

S-J

, et al. Morphing origami block for lightweight reconfigurable system. IEEE Trans Robot, 2020; 37(2):494–505.

57.

Xia

, Wang

, Yin

, et al. Design and mechanics of a composite wave-driven soft robotic fin for biomimetic amphibious robot. J Bionic Eng, 2023; 20:934–952.

58.

Patel

, Huang

, Luo

, et al. Highly dynamic bistable soft actuator for reconfigurable multimodal soft robots. Adv Mater Technol, 2023; 8(2):2201259.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

8.73 MB

4.75 MB

3.50 MB

18.87 MB

4.56 MB

6.89 MB

13.38 MB

18.53 MB

15.61 MB

18.06 MB

9.71 MB

25.40 MB

5.47 MB

3.93 MB

0.00 MB