Reticular Origami Soft Robotic Gripper for Shape-Adaptive and Bistable Rapid Grasping

Abstract

The top–down approach in designing and fabricating origami robots could achieve far more complicated functions with compliant and elegant designs than traditional robots. This study presents the design, fabrication, and testing of a reticular origami soft robotic gripper that could adapt to the shape of the grasping subject and grasp the subject within 80 ms from the trigger instance. A sensing mechanism consisting of the resistive pressure sensor array and flexible elongation sensor is designed to validate further the shape-adaptive grasping capability and model the rough shape and size of the subject. The grasping test on various objects with different shapes, surface textures, sizes, and living animals further validates the excellent grasping capabilities of the gripper. The gripper could be either actively triggered by actuation or passively triggered by a minimum of 0.0014 J disturbance energy. Such features make it particularly suitable for applications such as capturing underwater creatures and illegal drone control.

Introduction

In the natural environment, many animals and plants have processed the capability of rapid grasping in the generations of evolution due to the need for nutrition. One of the most famous examples is the Venus flytrap, which could close its upper and lower leaf within 0.3 s from the trigger instance to capture the insects that trigger the electrical stimulation between the lobes and midrib.¹ Sea anemones can also rapidly grasp prey with their membrane and retractor muscles.² The Bobbit worm (Eunice Aphroditois), which usually stays under the sand with an open jaw on the surface, could rapidly grasp the passing fishes and drag them down the sand,³ is also an excellent example of rapid grasping in nature.

The robotic grippers designed based on the rapid grasping mechanism of these plants and animals could achieve extraordinary capability in different application scenarios. The sensing–grasping integrated bistable gripper designed by Qi et al.² based on the structure of a sea anemone inspires further universal gripper design. The gripper designed based on the tentacles of sea anemone⁴ could also achieve higher adaptability than traditional gripper. The bistable gripper inspired by the Venus flytrap⁵ also shows an excellent pull-out force, which induces secure grasping.

Rapid grasping capabilities have been crucial in designing the robotic gripper for many applications. For example, illegal drones have been causing an increasing number of problems globally, and there are cases wherein drones cause large airports to stop for hours.⁶ Except for electromagnetic interference of the drone signal, a gripper that could capture the fast-flying drones could also be an effective way to prevent chaos in the airport or important facilities caused by illegal drones. Furthermore, ball capturing by grippers in the sports industry is usually achieved by complex motion tracking and predicting system,⁷ which is expensive to implement. In the cases mentioned above, bioinspired approaches for rapid gripper design could provide a simple and more elegant solution.

Furthermore, the studies show that 91% of the species in the ocean still need to be described and specified.⁸ The studies on marine biology would contribute greatly to the biodiversity of the ecosystem. However, the sublethal effect, which implies nonlethal harm to marine life, such as injury and acute stress response caused by capturing, releasing, and escaping, has threatened biodiversity.⁹ To avoid the potential sublethal effect caused by the traditional bycatch method during the study of marine life, a targeted, rapid, and nonharmful device that could be applied to capture marine life for scientific research is urgently needed.

The fluid elastomer actuator (FEA) is the most widely used soft actuation technology for soft grasping.¹⁰ The grasping is achieved with different motions, including bending^11–13 and twisting,^14,15 by pressurizing or vacuuming a chamber made with deformable materials with a constraining layer.¹⁶ In particular, Sinatra et al.¹⁷ developed an ultragentle soft robotic gripper for capturing specimens of gelatinous marine life with Young's modulus of 0.34 to 1.2 kPa, and the result showed that the jellyfishes used for the experiment show no sublethal effect after being grasped. Navas et al.¹⁸ designed a soft gripper with three single-channel diaphragm-type actuators.

The gripper was installed on a manipulator to verify its capabilities to harvest and manipulate different fruits. The low fabrication cost and robustness¹⁰ make the FEA gripper an ideal choice in many applications. However, the need for pressurizing devices significantly restricts the application of the FEA. For instance, the pneumatic and hydraulic structures generally require a carbon dioxide cartridge,¹⁹ gear pump,²⁰ and sometimes even a commercial air compressor²¹ for a larger actuation force, which is nonideal to be installed on a manipulator or remotely operated vehicles (ROVs).

An alternative approach is the contact-driven deformation gripper,¹⁰ which utilizes compliance material and mechanical structures^22,23 or metamaterial properties such as origami and kirigami structures^24–26 to conform to the contact surface of the grasping subject during the grasping to achieve soft grasping. A gripper inspired by the Fin Ray^® effect was tested to outperform the traditional gripper by 40%.²² The innovative jamming gripper²⁷ uses a single mass of granular material to conform and adapt to the shape of the grasping subject and a vacuum to harden the granular material to trap the subject in the gripper. The gripper can grasp objects that are difficult to be handled with traditional grippers, such as raw eggs. However, it is difficult for the jamming gripper to handle soft objects.²⁸

Other soft grasping approaches, such as shape memory polymers, have extremely low energy efficiency. Nylon and polyethylene have an efficiency of 1.08–1.32% and are even lower for magnet-driven actuators.²⁹ The high response time is also problematic. The gecko-inspired adhesive gripper requires a sophisticated fabrication process³⁰ and is sensitive to the cleanness of the environment.

In contrast, the sensing feedback from the gripper is of great importance to achieving precise and closed-loop control. The EGaIn sensor,³¹ based on the change in liquid metal resistance during deformation, could be an ideal choice to be embedded into a soft robotic gripper to measure the deformation. The textile silicone hybrid sensor³² can also acquire similar measurements based on the resistance change. Owing to its capacitor behavior, some dielectric elastomers actuated grippers can also function as a sensor with signal modulation techniques. The sensing accuracy is measured similarly to a laser displacement sensor.³³ The flexible pressure sensor³⁴ could also be an ideal choice for applications that require feedback control, such as crop harvesting.

To address the problems mentioned above, which can be concluded as the limitations on the adaptability to different shapes, surfaces, materials for secure and nondamaged grasping, and rapid grasping with compact design, we present a gripper capable of adapting the shape of the grasping object and rapid grasping (Supplementary Fig. S5). Rapid grasping can be triggered by a minimum of 0.0014 J disturbance energy applied on the gripper or actively by the linear actuator on the gripper and closing the gripper within 80 ms from the trigger instance. A sensing system is designed based on the customized carbon nanotube (CNT) elongation sensor and pressure sensor to model the shape of the grasping object.

Materials and Methods

Design and fabrication of the reticular origami gripper

As shown in Supplementary Figure S4A, a single piece of the gripper was cut out by a customized cutter die from a polypropylene (PP) film with a thickness of 10 μm. After cutting, 3M^® 9448 double-sided tape was applied to the film with two different patterns. Next, the PP films with different double-sided tape patterns were interlaced to form the reticular origami structure. Finally, a circular pattern is formed by stretching the structure and connecting the structure end-to-end.

Based on the two features of the gripper mentioned above, a thread-driven actuation structure is designed to enable the semiautomatic and automatic control of the gripper (Supplementary Fig. S2). A total of eight nylon fiber threads connect the top side of the gripper to a central reel through intermediate pulleys. A TowerPro MG996R continuous servo motor with a torque of 15 Kg cm drives the central reel through a timing belt with a gear ratio of 1:6.

A total of eight Adafruit^® N20 geared DC motors (150 rpm) connected to rubber wheels are placed in contact with the pulleys at the exit of the threads to keep the threads always tight within the actuation structure to avoid unexpected thread winding by rotating slightly faster than the central reel, while the threads were prereleased for the rapid grasping mode. In addition, a linear actuator is connected to a three-dimensional (3D)-printed cap to limit the bottom side of the gripper and trigger rapid grasping by pulling the cap downward.

The actuation structure was designed by SolidWorks, and the parts were sliced by Materialise Magics and printed by ZRapid SLA550 Stereo Lithography Apparatus (SLA) 3D printer. Suzhou Zhongrui Zhichuang 3D Technology Co., Ltd provided the white photosensitive resin.

Grasping force and range experiment

The grasping range measured by motion tracking shown in Figure 2A is acquired through the NOKOV (MARS4H, China) motion tracking device. The illustration of the experiment measuring the grasping force and range is shown in Supplementary Figure S1. The gripper was kept fixed on the base of the frame. A linear actuator controlled by a stepper motor was placed on top of the gripper and connected to the testing subjects through a thread. A load cell (JC-50, Leqing Jingcheng Instrument Co., Ltd.) is placed between the grasping subjects and records the real-time force measurement.

The effect of gravity is eliminated through precalibration before each measurement. The gripper initially grasped the objects, and the grasping force is recorded while the linear actuator pulls the testing object at constant velocity. The measurement data were analyzed and visualized by MATLAB (The MathWorks, Inc.).

Sensor design

The flexible elongation sensor³⁵ was fabricated by Multiwall CNTs (Nanocyl NC7000) with an average length of 1.5 μm. They were added to silicone elastomer (Ecoflex 0030; Smooth-On, Inc.) base and mixed for 5 min using a planetary centrifugal mixer (AR-100; Thinky). Subsequently, a certain amount of curing agent (base-to-curing agent ratio = 1:1) was added to the obtained solution and mixed for 2 min. After that, the CNT/silicone mixture (CNT mass percentage = 2.5 wt%) obtained was carefully injected into a stretchable rubber tube with a diameter of 1.5 mm. Then, the two ends of the rubber tube were sealed and heated at 60℃ for 30 min to fabricate stretchable conductive CNT/silicone composite fibers.

Finally, the stretchable and flexible strain sensor was prepared by attaching copper wires to the two ends of a CNT/silicone composite fiber as external electrodes. Note that the electrodes were covered with soft silicone adhesives (Sil-Poxy; Smooth-On, Inc.) to prevent mechanical failure during deformation. The sensor's elongation–resistance relationship is measured and calibrated before installation. The DFROBOT RP-C5-LT-LF5 thin film pressure sensor was attached to the gripper with a 3D-printed clip. The pressure–resistance relationship was calibrated through the datasheet provided by the manufacturer.

Two Arduino Mega 2560 Rev3 microcontrollers were used for the control and sensor data acquisition of the gripper. MATLAB Support Package for Arduino Hardware was used for real-time data acquisition and vitalization (Supplementary Fig. S3).

Rapid grasping

The MATLAB App DLTdv8a has been used to estimate the speed based on the recordings. The estimation was achieved by tracking a certain point on the moving object in the video. The pixel distance can be transferred into metric size by measuring the metric size of the reference object with the same distance to the camera. Once the change in metric position is calculated, the speed can be estimated given the 120-fps frame rate of the GoPro Hero video recording. The camera was placed in parallel with the ball falling trajectory in the experiment to ensure a relatively accurate two-dimensional (2D) estimation of 3D displacement.

In the minimum trigger kinematic energy experiment, the weight is still in the freefall state while hitting the gripper. Since the distance of the weight to the gripper is small, the kinematic energy when the weight hits the gripper can be calculated by $K E = m g x,$

where g denotes the standard acceleration of gravity, m is the weight mass, and x is the vertical distance between the initial height and the gripper. The Institutional Animal Care and Use Committee at Peking University authorized all experimental procedures and the care of animals.

Results

Origami soft gripper design and actuation

In the study, we presented a universal soft gripper inspired by the ancient art of paper cutting and folding, origami.³⁶ The origami structure enables the contact-driven deformation to adapt to the shape of the grasping subject and rapid grasping through the stored potential energy. A total of 152 pieces of quarter-circle shape PP films with a thickness of 10 μm were used to form a reticular pattern. The structure has a compressed width of 6 cm and could be stretched to >26 times its compressed length, as shown in Supplementary Figure S4 and Supplementary Movie S1. Such structure allows the gripper to grasp objects with a wide range of lengths and self-adapt to the shape and size of the object.

The origami structure was then connected end to end to form a circular pattern. The structure is opened by applying force along one of the ends. The origami structures close to the stretching end will deform and store the elastic potential energy, and the other end will be compressed due to the circular pattern, as shown in Figure 1A. Unlike the traditional contact-driven deformation soft gripper with two or more fingers, the circular shape of the gripper allows the gripper to wrap around the subject to create a uniform distribution of the contact stress for secure and gentle grasping.

FIG. 1.

The design and operating mode of the gripper. (A) The renderings of the gripper compared with the Venus Flytrap. (B) The illustration of the origami structure pattern that forms the circular gripper with the result of finite element method simulation. (C) The flow chart of adaptive grasping mode, in which the orange arrows indicate the actuation force, and the blue arrows indicate the force to the objects generated by the elastic potential energy. (D) The illustration of the potential application scenario of the adaptive grasping mode. The gripper is used to grasp and sort the tennis ball on the assembly line. (E) The flow chart of fast-grasping mode, in which the orange arrows indicate the actuation force, and the blue arrows indicate the force to the objects generated by the elastic potential energy. (F) The illustration of the potential application scenario of the fast-grasping mode. The gripper is used to capture a fast-moving drone. (G) The flow chart of the fast-grasping mode.

Another fascinating feature of the gripper is bistable fast grasping. As the actuation opens the gripper to a stable state, the elastic potential energy is stored internally through the deformation of the origami structure. The stable state could be broken by actuation or external disturbance, the potential energy will quickly be released, and the gripper will close within 80 ms to capture rapid moving objects.

Through the actuation structure shown in Figure 1B, the gripper can operate in two modes: self-adaptive grasping and rapid grasping. As shown in Figure 1C, the self-adapting grasping uses a central reel to control the length of threads to open and close the gripper. The reticular origami structure would adapt to the shape of the grasping subject. In the rapid grasping mechanism shown in Figure 1E, the process unfolds through a series of controlled actions utilizing threads and a linear actuator to manage the gripper's states. Initially, the threads engage to drive the gripper slightly past a threshold state, while concurrently, the linear actuator sustains this state, ensuring stability.

Once the gripper attains a stable state, the threads are released, permitting the gripper to close. This rapid grasping action can be initiated in two ways: by responding to an external disturbance or through a deliberate activation where the linear actuator presses downward on the gripper's center, pushing it beyond the threshold state to induce closure. The stored elastic energy will instantly release to close the gripper rapidly within 80 ms (Fig. 1G; Supplementary Movie S2).

Adaptive grasping force and range

The motion tracking of the closing end of the gripper first determines the grasping range. It is shown in Figure 2A that the grasping range is similar to a hemisphere with the radius of the distance of the end-to-end length of the gripper.

FIG. 2.

The grasping range and force, and the gripping sensor experiment. (A) The grasping range of the gripper acquired by optical motion tracking. (B) The plot of the relationship between the maximum grasping force and sizes of grasping subjects with different shapes and textures. (C) The adaptability of the origami gripper. (D) The design and placement of force sensors and elongation sensors on the gripper. (E) The results of the measurement of circular shape subjects with different sizes. (F) The results of the measurement of subjects with different shapes and sizes.

To further test the gripper's grasping range and capabilities of objects with different shapes and materials, foam and acrylic ball with six different diameters ranging from 3 to 25 cm, and the foam and acrylic cube with four side lengths ranging from 3 to 15 cm were used to measure the change in the grasping force during the grasping of a single object and determine the maximum grasping force.

It is shown in Figure 2B that when grasping a ball, the diameter of the ball that could generate the largest grasping force is 15 cm for a soft foam ball and 20 cm for the acrylic ball, a size smaller or larger than the optimal size would result in the decrease in the grasping force. The difference between the acrylic and foam balls, as well as the deviation of each ball, is small at the minimal and large diameter (0.22 N difference on average at 3 cm diameter, 0.07 N at 25 cm, and 3.1 N at 15 cm) and gradually increase in between the minimal and large diameter, the largest difference happens at 15 cm diameter (3.1 N).

The cube, in contrast, shows a more unstable size–force relationship. It is shown in Figure 2B that the grasping force is unstable for both soft and hard cubes in the first three sizes, with a significant increase and then a decrease at the size of 5 to 10 cm, which is caused by the corners of the cube interfering with the reticular structure.

As the size of the cube increases to 15 cm, the reticular structures will no longer interfere with the cube corners. Then the grasping force increases again to ∼10 N. In general, the soft cube's grasping force exceeds that of the hard cube. The difference is minimal at extremely small and large side lengths, but more pronounced in between. Specifically, there is a difference of 1.2 N at 3 cm side length, 0.9 N at 15 cm, and a notable 3.7 N at 5 cm.

The maximum grasping force is always larger for the cube compared with the ball since for the ball and cube with the same diameter and side length, the cube's perimeter is $\frac{4}{π}$ times larger than that of the ball. The grasping force depends on the elastic potential energy of the origami structure. A smaller perimeter would cause smaller deformation in the structure and, thus, a smaller grasping force. The decrease in the grasping force at a large size can be explained by the change in the direction of the force as the gripper is being opened.

As shown in Supplementary Figure S6, the horizontal hold, which could be considered the gripper's holding force, decreased when the gripper was opened to a certain level. As the grasping force increases, the force differential between the acrylic and foam also increases, due to the surface friction of the materials. The foam could cause significantly larger friction than the smooth acrylic surface.

Shape adaptability experiment and sensors design

The adaptability enabled by contact-driven deformation of the reticular origami structure is essential for many applications. The gripper could perfectly adapt to various shapes, as shown in Figure 2C. However, the high flexibility and deformation usually indicate high difficulty in sensor deployment measurement. Therefore, an integrated sensing system with flexible pressure and elongation sensors is designed to enable feedback control and explore the shape-adapting capabilities of the gripper. As shown in Figure 2D, 12 thin film pressure sensors are placed on the top of the gripper, where most contact happens during gripping. A customized elongation sensor is placed on the middle part of the gripper, where the largest elongation happens during grasping.

The reliability of the elongation sensor was first verified by grasping objects of different sizes in a circular shape, including basketball, aluminum can, heavy-duty tape, transparent box, bowls, and tennis balls (Supplementary Movie S3). The measured length is obtained by the linear relationship between the resistance and elongation length and the linear relationship between the middle and top perimeter of the gripper. Comparison between the actual and the measured length is shown in Table 1. The error is as small as 0.8 cm (1.6%) when measuring objects with large diameter such as basketball, and the error increases exponentially as the size of the measuring objects gets smaller. (4 cm and 32.3% for a tennis ball.)

Table 1.

The Elongation Sensor Length Measurement Error

Object	Measured length (cm)	Actual length (cm)	Error (cm)
Basketball	48.9	49.7	0.8
Aluminum can	29.3	27	2.3
Transparent box	24.7	20.1	4.6
Tennis ball	8.4	12.4	4

The foam in the shape of the equilateral triangle, right triangle, square, and rectangle (Fig. 2F) was used to test the integrated sensing system and verify the shape-adapting capabilities. The real-time measurements are exported with the serial ports on the microcontroller and visualized through MATLAB. The virtualization of the measurement result is shown in Figure 2E and F.

Fast grasping experiment

As described in the previous paragraph, rapid grasping can be triggered by linear actuation or an external disturbance that drives the gripper over the threshold position. First, to test the rapid gripping triggered by external disturbance, the linear actuation is disabled. Then, a shuttlecock was dropped from a height of 1 m to different positions shown in Figure 3A on the gripper at the thresholding state to test the gripper's rapid grasping capability triggered by external disturbance without actuation. As shown in Supplementary Movie S4, in all the testing, the rapid grasping was triggered by the shuttlecock in all the positions on the gripper, and the velocity at the instant that the shuttlecock touched the gripper was estimated to be 3.23 m/s.

FIG. 3.

The experiments on the minimum disturbance energy could trigger the fast-grasping mode. (A) The illustrations of different landing spots of the shuttlecock on the gripper trigger the active fast grasping. (B) The relationship between the contacting spot on the gripper and the minimum height of the weight could trigger the fast grasping for three different weights.

An experiment was designed to determine further the minimum disturbance energy that could trigger the rapid grasping on different positions on the gripper. The weights of 5, 10, and 20 g were dropped repeatedly from different heights onto different positions on the gripper to determine the minimum height of dropping the weight that could trigger the rapid grasping in different positions.

The result shows that the center of the gripper requires the minimum disturbance kinematic energy to be triggered (0.0014 J). In contrast, the middle area requires the largest disturbance kinematic energy (0.0035 J), which could be explained by the quarter-circular shape of the PP films that form the origami structure. The disturbance in the edge area is essentially amplified by the Leverage effect, which drives the structure over the thresholding state and triggers the rapid grasping. In general, the passive grasping of the gripper is extremely sensitive and requires as little as 0.0014 J kinematic energy to be triggered.

Real-life scenario rapid grasping and underwater experiment

The main purpose of implementing adaptive grasping is that, in real-life applications, the variation of shape, size, surface material, and stiffness is much larger than in a well-controlled factory or laboratory environment. Hence, to further validate the design reliability, a wide range of commonly seen objects in daily life were used to test the adaptive grasping capability, including a large rubber ball, gaming controller, toy drone, small plastic ball, three ping-pong balls, badminton, bowl, sunglasses, tissue box, paper cup, light bulb, plastic bottle, boxing glove, nylon threads spool, and plush toy (Supplementary Movie S5).

The objects have relatively large variations in size, shape, surface material, and stiffness. The gripper could grasp all objects reliably with more than a 95% success rate, which shows high adaptability (Fig. 4A). The grasping enabled by flexible wrapping could pick up multiple objects simultaneously and pick up objects from different directions (Fig. 4B). None of the objects shows permanent deformation or damage after being grasped.

FIG. 4.

The gripper's capability of grasping real-life objects. (A) The flow chart of the procedure of the grasping experiment. (B) The illustration of the gripper picks up a wide range of real-life objects with different sizes, shapes, and surface textures.

As shown in Figure 5A, a toy drone was controlled to land vertically on the gripper, land on the gripper with specific angles, and fly horizontally over the gripper (Supplementary Movie S6). The passive and adaptive grasping models could perfectly collaborate to capture the drone in all the scenarios. The gripper could even grasp much smaller and lighter objects with its active rapid grasping mode, as shown in Figure 5B and Supplementary Movie S7. A ping-pong ball was served from one end of the table toward the gripper's grasping range on the opposite end.

FIG. 5.

The experiments of the gripper's fast-grasping mode. (A) The illustration of the gripper capturing a drone that (1) lands vertically onto the gripper, (2) lands at a certain angle onto the gripper, and (3) flies horizontally over the gripper's grasping range. (B) The illustration of the gripper capturing a ping-pong ball that was flying over the grasping range. (C) The illustration of the gripper capturing ball and badminton with 45° and 90° angles. (D, E) The illustration of the gripper capturing a ball that has been thrown up with 180° angle (upside down), and a goldfish underwater.

With its rapid actuation, the gripper successfully and accurately captured the ball. All the current noncontact grasping, such as the case in which a drone flies over the gripper, is controlled manually through visual feedback to demonstrate the possibility of utilizing a camera or time of flight (ToF) sensor to enable noncontact grasping.

To test whether the rapid grasping is angle dependent, the gripper is placed at 45°, 90°, and 180° upside down to grasp a plastic ball, a shuttlecock, and a rubber ball, as shown in Figure 5C, D, and Supplementary Movie S8. In all cases, the gripper could securely capture the object with active and passive rapid grasping. The feature could enable the gripper to be used under the water to capture marine lives that are agile under the water and have a relatively soft body from different positions and angles.

Figure 5E and Supplementary Movie S9 show that the gripper was used to capture a goldfish under the water, whereas the space caused by the curvature shape of the gripper could protect the goldfish from physical harm. None of the tested animals showed a change in behavior after the experiment.

Discussion and Conclusion

In general, the origami gripper shows an excellent adaptive grasping capability. It can grasp a ball with a diameter ranging from 3 to 25 cm and a cube with a side length ranging from 3 to 15 cm. The sensing system is designed based on contact-driven adaptive grasping and shows good precision in sensing the perimeter and shape of the grasping object. The rapid grasping, which could close the gripper within 80 ms from the trigger instance, can be triggered by a minimum of 0.0014 J disturbance kinematic energy or the linear actuator placed on the gripper.

Adaptive grasping

The gripper achieves active adaptive grasping through the compliant origami structure and the potential elastic energy caused by deformation. The experimental result shows that the active grasping force is perimeter and surface material dependent. A small perimeter would cause less deformation on the gripper and, thus, less grasping force. However, as the gripper's grasping is a circular-like motion around the center, the direction of the force would change when the perimeter of the grasping object exceeds the grasping range, which causes the decrease in horizontal hold force.

This means that the size of the grasping object should fit the size of gripper to generate the largest adaptive grasping force to hold the object securely. Further exploration can be made to determine the relationship between the PP film thickness, gripper size, object size, and gripper curvature for the future application of the gripper. The surface material is also important to generate higher friction between the object and gripper during the grasping.

The fillet radius of the corner of the PP film could be further adjusted to avoid shape contact, especially for the application of marine life capturing. An object with a sharp corner, such as the cube in the grasping force experiment, could sometimes interfere with the reticular structure of the gripper. Future improvements can be made to find the soft stretchable material to cover the edge as well as the inner surface of the gripper to enable softer and safer contact between the extremely fragile object and the object with an irregular shape.

Shape adaptivity and sensor design

A sensing system consisting of an elongation sensor and pressure sensor array distributed on the edge of the gripper was designed to measure the perimeter and shape of the grasping object to identify the rough shape of the object and enable close-loop control of the gripper. In addition, real-time visualization of the measurement user interface (UI) is designed using MATLAB. The result shows the satisfactory size and shape recognition of various objects.

Further improvements can be made to increase the resolution of the pressure sensor array. The current sensor array comprises 12 thin film sensors. This setup results in notable blind spots during wide openings of the gripper as the distance between each increase, especially when measuring objects with angular shapes or cavities. If the edge or cavity falls within these blind spots, measurement precision could be compromised.

The current UI displays the pressure sensor measurement through the color on the corresponding section on a circle. If the resolution could be increased, the UI can be upgraded to re-establish the geometric shape of the measured object based on the perimetric and pressure sensor data. The fascinating feature can be used to create the 3D model of the grasping object through repeated measurement on different height levels on a single object, like a 3D scanner. Furthermore, the sensors enable close-loop precise control of the gripper for safe contact between the gripper and the object.

Fast grasping

The rapid grasping mode can be triggered by linear actuation and external disturbance. Different regions within the gripper require different disturbance energy to be triggered. The experiment result indicates that the outer edge is the most sensitive area, which requires barely 0.0014 J disturbance energy to be triggered. In contrast, the area between the gripper's outer and inner edges is the least sensitive area. However, even for the most insensitive areas, the disturbance energy required to trigger rapid grasping is extremely small (0.12 J), which means a shuttlecock dropped from 1 m height is more than enough to trigger the grasping.

In contrast, employing linear actuation can effectively mitigate the primary drawback associated with active rapid grasping: the necessity of introducing an external disturbance on the gripper to initiate the grasp. With the linear actuation, the gripper can capture a ping-pong ball or a drone that horizontally flies over the grasping range. Although linear actuation is currently manually controlled by humans with visual feedback, it is sufficient to demonstrate the possibility of future automatic control.

Most importantly, the rapid grasping is not angle or position dependent, which enables the gripper to capture the object at different angles and positions. Different sensors, such as infrared, ultrasonic proximity sensor, or computer vision algorithm, could be used in future work to activate the linear actuator with better timing to capture the object with even higher motion velocity. In the applications of capturing the live animal, a soft stretchable layer or larger curvature radius could be considered depending on the size and vulnerability of the animals to protect them.

Future work

The current prototype of the gripper is still in the proof of concept and characterization stage. The gripper's mechanical design must be improved to fit into different applications. In the current underwater experiments, only the reticular structure part was submerged, limiting the testing's degrees of freedom. In future tests, the thread-driven structure and linear actuator will be sealed in a waterproof chamber. The gripper will then be installed on the manipulator on an ROV to evaluate its underwater grasping capabilities. The gripper can also be modified to be placed on the robotics arm to test its performance in sports industry. In this scenario, a camera or ToF sensor must be installed and tested to enable noncontact grasping in a valid grasping range.

Data and Materials Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Footnotes

Author Disclosure Statement

The authors declare no conflict of interest.

Funding Information

This work was supported in part by the National Natural Science Foundation of China under Grant 12272008, Grant U23B2037, and Grant U22A2062 and in part by the Beijing Natural Science Foundation under Grant 3242003. Xingwen Zheng is a JSPS postdoctoral fellow. He is supported by the Japan Society for the Promotion of Science (JSPS, Grant No. 22F31357).

Supplementary Material

References

Volkov

, Adesina

, Markin

, Jovanov

. Kinetics and mechanism of dionaea muscipula trap closing. Plant Physiol, 2007; 146(2):323–324; doi: 10.1104/pp.107.108241.

, Xiang

, Ho

, Rossiter

. A sea-anemone-inspired, multifunctional, Bistable Gripper. Soft Robot, 2022; 9(6):1040–1051; doi: 10.1089/soro.2020.0176.

Fellowes

, Barnett

, Ackland-Snow

30-Second Zoology: The 50 Most Fundamental Categories and Concepts from the Study of Animal Life. Ivy Press: London; 2020.

Yang

, Ren

, Yang

, et al. Design of a flexible capture mechanism inspired by sea anemone for non-cooperative targets. Chinese J Mech Eng, 2021; 34(1):77; doi: 10.1186/s10033-021-00594-z.

Wang

, Khara

, Chen

. A soft pneumatic bistable reinforced actuator bioinspired by Venus Flytrap with enhanced grasping capability. Bioinspir Biomim, 2020; 15(5):056017; doi: 10.1088/1748-3190/aba091.

Singh

Throwback: When drone sightings caused chaos at London gatwick. 2022. Available from: https://simpleflying.com/gatwick-drone-sightings-recap/ [Last accessed: August 31, 2023].

Rapp

HH.

A ping-pong ball catching and juggling robot: A real-time framework for vision guided acting of an industrial robot arm. The 5th International Conference on Automation, Robotics and Applications;, 2011; doi: 10.1109/icara.2011.6144922.

Mora

, Tittensor

, Adl

, et al. How many species are there on Earth and in the ocean?. PLoS Biol, 2011; 9(8):e1001127; doi: 10.1371/journal.pbio.1001127.

Wilson

, Raby

, Burnett

, et al. Looking beyond the mortality of bycatch: Sublethal effects of incidental capture on Marine Animals. Biol Conservat, 2014; 171:61–72; doi: 10.1016/j.biocon.2014.01.020.

10.

Shintake

, Cacucciolo

, Floreano

, Shea

. Soft robotic grippers. Adv Mater, 2018; 30(29):1707035; doi: 10.1002/adma.201707035.

11.

Deimel

, Brock

. A novel type of compliant and underactuated robotic hand for dexterous grasping. Int J Robot Res, 2015; 35(1–3):161–185; doi: 10.1177/0278364915592961.

12.

Tawk

, Gillett

, in het Panhuis

, Spinks

, Alici

. A 3D-printed omni-purpose soft gripper. IEEE Trans Robot, 2019; 35(5):1268–1275; doi: 10.1109/tro.2019.2924386.

13.

Wang

, Or

, Hirai

. A dual-mode soft gripper for food packaging. Robot Auton Syst, 2020; 125:103427; doi: 10.1016/j.robot.2020.103427.

14.

Yang

, Cooper

, Liu

, et al. Twining plant inspired pneumatic soft robotic spiral gripper with a fiber optic twisting sensor. Opt Expr, 2020; 28(23):35158; doi: 10.1364/oe.408910.

15.

Lee

, Wang

, Zheng

. Twister hand: Underactuated Robotic Gripper inspired by Origami Twisted Tower. IEEE Trans Robot, 2020; 36(2):488–500; doi: 10.1109/tro.2019.2956870.

16.

Gorissen

, Reynaerts

, Konishi

, et al. Elastic inflatable actuators for soft robotic applications. Adv Mater, 2017; 29(43):1604977; doi: 10.1002/adma.201604977.

17.

Sinatra

, Teeple

, Vogt

, et al. Ultragentle manipulation of delicate structures using a soft robotic gripper. Sci Robot, 2019; 21:4(33); doi: 10.1126/scirobotics.aax5425.

18.

Navas

, Fernandez

, Sepulveda

, et al. Soft gripper for robotic harvesting in precision agriculture applications. 2021 IEEE International Conference on Autonomous Robot Systems and Competitions (ICARSC);, 2021; doi: 10.1109/icarsc52212.2021.9429797.

19.

Marchese

, Onal

, Rus

. Autonomous soft robotic fish capable of escape maneuvers using Fluidic Elastomer actuators. Soft Robot, 2014; 1(1):75–87; doi: 10.1089/soro.2013.0009.

20.

Katzschmann

, DelPreto

, MacCurdy

, Rus

. Exploration of underwater life with an acoustically controlled soft robotic fish. Sci Robot, 2018; 3(16):eaar3449; doi: 10.1126/scirobotics.aar3449.

21.

Kim

, Cha

. Soft pneumatic gripper with a tendon-driven soft origami pump. Front Bioeng Biotechnol, 2020; 8:461; doi: 10.3389/fbioe.2020.00461.

22.

Crooks

, Vukasin

, O'Sullivan

, et al. Fin Ray^® effect inspired soft robotic gripper: From the RoboSoft Grand Challenge toward optimization. Front Robot AI, 2016; 3:70; doi: 10.3389/frobt.2016.00070.

23.

Liu

C-H

, Huang

G-F

, Chiu

C-H

, Pai

T-Y

. Topology synthesis and optimal design of an adaptive compliant gripper to maximize output displacement. J Intell Robot Syst, 2017; 90(3–4):287–304; doi: 10.1007/s10846-017-0671-x.

24.

, Stampfli

, Xu

, et al. A vacuum-driven origami “magic-ball” soft gripper. 2019 International Conference on Robotics and Automation (ICRA);, 2019; doi: 10.1109/icra.2019.8794068.

25.

Liu

, Wohlever

, Ou

, et al. Shake and take: Fast transformation of an origami gripper. IEEE Trans Robot, 2022; 38(1):491–506; doi: 10.1109/tro.2021.3076563.

26.

Yang

, Vella

, Holmes

. Grasping with Kirigami shells. Sci Robot, 2021; 6(54):eabd6426; doi: 10.1126/scirobotics.abd6426.

27.

Brown

, Rodenberg

, Amend

, et al. Universal robotic gripper based on the jamming of granular material. Proc Natl Acad Sci USA, 2010; 107(44):18809–18814; doi: 10.1073/pnas.1003250107.

28.

Zhakypov

, Heremans

, Billard

, Paik

. An origami-inspired reconfigurable suction gripper for picking objects with variable shape and size. IEEE Robot Autom Lett, 2018; 3(4):2894–2901; doi: 10.1109/lra.2018.2847403.

29.

Boyraz

, Runge

, Raatz

. An overview of novel actuators for Soft Robotics. Actuators, 2018; 7(3):48; doi: 10.3390/act7030048.

30.

Glick

, Suresh

, Ruffatto

, et al. A soft robotic gripper with Gecko-inspired adhesive. IEEE Robot Autom Lett, 2018; 3(2):903–910; doi: 10.1109/lra.2018.2792688.

31.

Vogt

, Park

Y-L

, Wood

. Design and characterization of a soft multi-axis force sensor using embedded microfluidic channels. IEEE Sens J, 2013; 13(10):4056–4064; doi: 10.1109/jsen.2013.2272320.

32.

Atalay

, Sanchez

, Atalay

, et al. Batch fabrication of customizable silicone-textile composite capacitive strain sensors for human motion tracking. Adv Mater Technol, 2017; 2(9):1700136; doi: 10.1002/admt.201700136.

33.

Jung

, Kim

, Choi

. A self-sensing dielectric elastomer actuator. Sens Actuat A Phys, 2008; 143(2):343–351; doi: 10.1016/j.sna.2007.10.076.

34.

, Li

, Shi

, et al. Recent developments for flexible pressure sensors: A review. Micromachines, 2018; 9(11):580; doi: 10.3390/mi9110580.

35.

Tang

, Jia

, Wang

, et al. Highly stretchable core–sheath fibers via wet-spinning for wearable strain sensors. ACS Appl Mater Interf, 2018; 10(7):6624–6635; doi: 10.1021/acsami.7b18677.

36.

Rus

, Tolley

. Design, fabrication and control of Origami Robots. Nat Rev Mater, 2018; 3(6):101–112; doi: 10.1038/s41578-018-0009-8.

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.

59.61 MB

0.73 MB

0.00 MB