Soft Grippers in Robotics: Progress of Last 10 Years

Abstract

This paper is dedicated to soft grippers, robot tools with a wide application area in various activities where an accurate and delicate grabbing movement is required such as routine manipulation tasks with fragile objects, operation in unknown or dangerous environments, and manipulation with unknown shape objects, as well as exploring the depths of the sea or harvesting vegetables in agriculture. The main goal of this paper is to review and systematize the main ideas about and achievements of soft grippers published from 2015 to 2024. The paper provides a statistical analysis of the performed research and systematized advancements of soft grippers according to their operating principle, forces and effects that enable their operation, and the properties of potential manipulation objects. Grippers inspired by nature are also discussed, as most successful solutions are based on ideas derived from nature. This study discusses the latest achievements of soft grippers and their various applications and presents a unique distribution of soft grippers according to the physical principle of the forces they act on, according to the size of the object to be grasped, and according to technological realizations. The results of this analysis can be useful for practical gripper users aiming to improve their workplace and find optimal design solutions, for gripper manufacturers or developers, or for scientists of material sciences looking for applications for their products.

1. Introduction

Modern soft grippers are unique devices enabling robots’ interaction with the external environment in a certain actuating mode by utilizing the properties of flexible materials and complex structure and shape design inspired by nature [1,2].

Robotization has touched practically all areas of today’s industry, including applications such as agricultural harvesting, commodity sorting, and medical rehabilitation or targeted therapy, as well as environment cleaning [3,4,5,6,7].

Specific and delicate manipulation tasks such as fruit or berry harvesting or processing require developing robot tools with stiffness and control properties unachievable when using a traditional approach with rigid links and conventional actuators.

Soft grippers are one of the possible solutions to the challenges mentioned. They allow one to perform complicated tasks with fragile and soft components [8]. Design of soft robotic grippers is a rapidly growing field of engineering, with enchanting prototypes with the application of new smart materials for mechanical devices resulting in untethered soft robotic grippers with “embodied” intelligence for soft systems [9,10,11].

Grasping and manipulation are fundamental actions of complex living organisms, enabling their interaction with the environment. Different morphologies and grasping methods in nature have evolved to adapt to specific environmental conditions. For example, bird feet are composed of rigid frames and soft joints, and some plants can bend and construct capture points and have the ability to perform hydrostatic-based complex movements in 3D space [12]. Similarly, soft robotic grippers use different designs, natural forces, and effects to grab the object. Among the most common typologies are mechanical grippers with jaws, vacuum grippers, magnetic grippers, and anthropomorphic grippers [13].

Selection of the gripper mechanical structure and operating principle is typically based on the requirements raised by the manipulation task, which consider object geometry, stiffness, weight, required transfer speed, and trajectory, and a set of other related parameters depending on the individual use case. For example, an industrial robotic arm can achieve high acceleration in positioning and picking tasks; therefore, it is necessary to ensure sufficient force to lock an object in the gripper reliably. As some authors point out, robotic hands’ velocity and acceleration in typical manipulation tasks could be in the ranges from 2 m/s to 10 m/s and 10 m/s² to 100 m/s², respectively [14,15]. Such conditions require high friction between the object and gripper surfaces or geometrical locking to overcome the forces of inertia.

More delicate manipulation tasks, such as surgery and other medical or micromanipulation applications, require the use of advanced grippers with force control and sometimes even feedback for the operator [16], ensuring implementation of safety features, adaptive behavior, and comfortable communication with a robot [17]. For example, the master surgical robotic systems complemented with auxiliary surgical robots connected to the central console station allow information to be received directly from the operating site through haptic feedback [18]. Therefore, robotic grippers are vital not only for the grabbing action itself but also play a crucial role in the whole manipulation process, especially in intelligent manufacturing and medicine [19].

Soft grippers with artificial muscles based on the deformation of conductive polymers under electric impact [20] are some of the promising solutions, together with variable-stiffness grippers where differential gears and springs allow joint stiffness to be adjusted [21]. Robots’ ability to simulate human actions by characterizing objects by touching them is a current trend in modern robotics. The two directional signals and data flow between the microprocessor controlling the soft gripper and the robot operating system are essential for autonomous robotic systems [22].

Multifunctional grippers are another case gaining the attention of scientists and technicians. Machines for surgery, deactivation of explosives, and handling soft, fragile objects are common cases where soft grippers are superior compared to classic ones. Underwater grippers are a separate class of robot grippers. Existing deep-sea robotic manipulation solutions have historically been driven by the oil industry, resulting in destructive interactions with underwater life and the implementation of specific key technologies [23]. Growing demand for robotic grippers in the fields of medical operations, auxiliary maintenance, land rescue, underwater grasping, and space manipulation [23,24] fosters the development of new designs and implementation of new materials into soft grippers.

Soft grippers have different actuation techniques involving fluidic elastomer actuators, tendon-driven actuators, electroactive polymers, shape memory jamming, electromagnetic and stimuli-responsive gels, and polymers [25]. Such grippers often have ten or more degrees of freedom per actuator and include some components that passively conform the geometry of grasping objects, allowing them to grip objects with different shapes without additional adjustment [26]. The grasping of non-structural objects sensitive to contact force is an essential issue in the development of soft grippers [27]. L. Zhou [28] proposed to classify soft grippers by gripping method in the following categories: impactive (jaws, clamps, pinch devices), ingressive (pins, hooks, hackles), astrictive (vacuum suction, magneto- and electro-adhesion), or contigutive (chemo- and thermo-adhesion). Such classification explains the importance of contact quality, surface penetration, or another direct impact on the object.

In many cases, grippers might be tasked for pick-and-place operations without knowing the conditions of the environment. Different designs and techniques have been developed to increase the flexibility of grippers in unknown environments, such as using vision systems, sensory feedback, and novel mechanisms with flexibility in gripping [29,30,31].

The soft gripper can interact with its surroundings safely and flexibly because it comprises flexible materials. Certain technical systems have been developed through research on flexible grippers, but they are still in their infancy. The development of a universal gripper with the ability to grasp unfamiliar objects with unpredictable softness, surface properties, and shape is still challenging [32].

Various tools for manipulating fragile objects are the main topic of this paper. Most of the attention is paid to nature-inspired ideas and sustainable solutions as intertwined and inseparable topics. Also, the most attractive ideas for gripper designs and the scope of applications of the last decade are represented here. This review focuses on the progress of research on soft grippers and the efficient grasping of unknown objects with a complicated structure.

This paper aims to systematize the announced soft grippers according to their mode of operation, the force of the holding of the object, the principle of gripper operation, and their combinations. The uniqueness of this study is not only the analysis of the latest achievements of soft grippers and their various applications in agriculture, medicine, and industry but also the presentation of a unique classification of soft grippers according to the physical principle of operation and method used to hold the object, as well as the size of the object to be grasped. This classification will help manage a huge information flow and help other authors find a place for the products they create. Furthermore, this is relevant for gripper developers and developers of new materials and polymers looking for applications for their products.

2. Methodology of Research

This research analyzes and integrates interdisciplinary data from publications on soft gripper research, manufacturing, and mechanical properties investigation. Information was searched for in ScienceDirect, Google Scholar, and IEEExplore databases using the following keywords: geometric fixation, frictional forces, soft grippers for underwater, soft grippers for agriculture, soft grippers applied in surgery, soft grippers for medicine adhesion force in soft grippers, electrostatic adhesion in soft grippers, friction forces soft grippers, magnetic and electromagnetic grippers, pneumatic soft grippers, vacuum soft grippers, water-hydraulic grippers, micrometer size grasping objects, millimeter size grasping objects, grippers for fragile objects, grippers for unknown geometry of grasping objects, grippers for soft and sensitive grasping objects, nature-inspired soft grippers.

The publications included in this review were selected if they had clear, detailed pictures, referenced possible applications, tested in practice, and provided original solutions defined in detail. Attention was paid to the most famous authors and publications in the most popular journals and conference materials where the researchers prefer to present their technical solutions rather than fundamental extensive research.

At first, articles included in the analysis were sorted into three groups using the gripping method. These methods were the geometrical locking of the object in the gripper, implementation of friction force to hold the object, and a combination of those two approaches.

Further, these three groups of grippers were sorted according to the technological solution and used actuation types. In every section of this article, nature-inspired grippers are mentioned to explain the main operation principles. Whereas the grasping procedure can be divided by the fragility of the object to be grabbed, the grippers are divided by grasping object geometry, size, fragility, and results, which are represented in tables in the text below. The practical application of grippers is also discussed since most grippers are designed for a specific activity or are classified as multifunctional/universal by their authors.

The analysis completes this study, and a structured table of examples summarizes the most recent advancements, defines research gaps in soft gripper development, and provides recommendations for further research.

3. Statistical Analysis of Included Papers

This chapter provides systematic statistical information about the publications that have appeared in the last ten years regarding soft gripper development and their main application fields.

Based on the ScienceDirect database, more than 1000 articles with the keyword “soft gripper” were published yearly. The number of publications grew yearly and reached about two thousand in 2024. Figure 1 represents the growth of the number of scientific papers in general and their distribution regarding the main operational principle. As seen from the data presented in Figure 1, the total number of publications on soft grippers is about 2000 per year, and the distribution of those papers according to the main operating principle is almost equal and reaches a few hundred in each group. The other publications are mainly dedicated to the mathematical modeling of complex material behavior or complex control solutions.

From the statistics in Figure 1, it can be seen that all types of operation principles of soft grippers are of similar interest to researchers. Such a situation points out that this area of robotics is still under development, and standard technological solutions do not satisfy market demand.

Figure 2 represents the distribution of collected data by keywords according to areas of implementation. Hence, surgery is part of medicine, but articles dedicated to medical purposes and surgery purposes have about the same number of publications, and they are solving different problems. The unknown environment is a very specific field of interest, and there were fewer than 10 publications on this topic per year, except in 2024, when 16 papers were published. Multifunctional soft grippers are in demand, and the number is growing every year, from 1 in 2017 to 49 in 2024.

As shown in Figure 2, there are different tendencies in comparing the growth of publication in different application fields. Some topics, like soft grippers for underwater and agriculture applications, experienced a steady increase, while topics like soft grippers for surgery, medicine, and industry developed unevenly, with a drop in the number of publications observed in some years, probably because of some technological limitations being faced.

The lower level of attention paid to the soft universal grippers compared to industrial ones can be explained by the different approaches developers use to develop the gripper concept. While industrial grippers have a known application and practical test base, soft universal grippers are limited by various factors.

Figure 3 represents the number of articles concerning actuation type and classifies them according to these keywords: hydraulic, electrostatic, pneumatic, vacuum, and magnetic working principle. Based on these data, it can be seen that magnetic soft grippers are the most popular research objects compared to others.

Hence, based on information from the ScienceDirect database, in 2023–2024, two thousand articles on soft grippers were published.

According to the prescription to the science field, the articles were distributed as follows: 38% of articles were in engineering, 25% in material science, 6% in computer sciences, 9.6% in medicine and Dentistry, and 5.5% in Physics and Astronomy. Less than 4% mentioned Biochemistry, Chemistry, Chemical Engineering, and agricultural, Earth and Planetary Sciences. The number of reviews dedicated to soft grippers in the analyzed period was less than a hundred.

According to the results of the analysis, the most challenging case in the applications for soft grippers is an operation in an unknown environment with fragile objects of unknown geometry. Almost all types of such activities require soft grippers to perform a delicate grasp of an object using tactile or other feedback and considering the environmental conditions.

4. Types of Gripper Systems

There have been various attempts and approaches to classify robotic grippers; however, due to non-uniform international regulations, in the sense of existing standards, and an enormous variety of grippers, there is no unified approach. Many scientists have made significant attempts to provide various classification methods. Z. Samadikhoshkho [33] proposed the classification of soft grippers based on configuration, actuation, application, size, and stiffness, gave structured information about gripper type (working via cable, vacuum, or pneumatic, hydraulic, or servo-electric mechanisms), and, at the same time, named the most effective other systematization methods. Sh. Zaidi [34] reviewed technologies for soft grippers and actuation principles and presented six major kinds of technologies that are either used independently for actuation or in combination: pneumatic actuation, vacuum actuation, cable-driven actuation, shape memory alloy actuation, electroactive polymer actuation, electro-adhesive actuation, and other types. Classification by gripping technologies, including actuation, stiffness control, and adhesion, proposed by J. Shitake [35], includes technologies, lifting ratio, gripper and object size, and response time. E. Navas reviewed grippers with agricultural applications [36] and systematized them according to the picking pattern and control philosophy.

4.1. Geometric Bonding Grippers

This section discusses the innovations of soft grippers that use geometrical bonding as a method to hold the object within the gripper, typically using movable or deformable fingers of a specific shape. Grippers of this type include most typical actuation techniques such as electromechanical, hydraulic, pneumatic, magnetic, adhesive, and electric techniques.

Electromechanical actuation is the most simple and typical approach of the grippers that use geometrical fixation to hold the object and have a structure with two–five movable fingers [33], where the mechanical compression of the gripper can be directly applied for simple tasks, such as tool exchange or another actuation [37] for automatically planned grasping. The optimal gripper is a three-point contact gripper, which guarantees the object’s safety and automatic centering. The soft fingers can deal with unknown, delicate objects without damaging them, as demonstrated by their ability to visualize a light bulb [38].

A tendon-driven passive structure gripper proposed by S. Terrile et al. [39] is an example of a typical electromechanical gripper. The gripper uses tactile sensing and twists objects to obtain a three-dimensional tactile-feedback-based model of an object.

The electromechanical two-finger gripper (Figure 4) proposed by Chia-Ju Peng [40] operates using the electromechanical properties of a polymer placed in the reservoir, where, under applied electric potential, the polymer network provides ionic bonding, ensuring mechanical strength of the material and causing its deformation, resulting in bending of the gripper finger. Here, the grab function is realized by employing complex internal structures and specific material properties.

W. Wang [25] proposed a soft gripper based on shape memory alloy. This gripper uses one deformable finger and additional sensors for stiffness control (Figure 5) when grasping deformable objects without recognizing them. The proposed gripper consists of a single soft finger, which is an SMA-based hinge actuator that can induce hinge-like bending deformation. Feedback about finger bending deformation in real time was realized using a cellphone camera 2D space. As a demonstration, soft planar grippers with the desired deformation were eventually used to grip deformable objects, including flowers. That task is challenging due to their complex shape, fragility, ability to bond into groups, and sensitivity to mechanical impact.

The grippers for fruit collecting working on geometrical principles have a wide variety of design solutions [41,42,43,44]. For example, a tactile information-collecting method to be used during fruit grasping is proposed by Y. Wei [45], resolving the geometric interaction between the gripper and the fruit is proposed by K. Fu [46], and tomato collecting improved by the sequence planning method in a multi-cluster environment based on multi-views is presented for tomato clusters by N. Dai [47]. A gripper with four fingers, employing a neural network and deep-learning-based detection method, is proposed by J. Liang [48]. The teardrop-shaped sheets, by controlling the moving frame position online, enabled them to grasp berries, as proposed by Mun [49]. E. Navas proposed a three-finger gripper made from flexible thermoplastic elastomer filament and evaluated Ogden’s theory using pneumatic actuation and a rotating rigid structure enabled for blueberry collection [50].

The six-rotor drone integrated with the finger soft gripper expands the possibilities of using the gripper proposed by Y. Zhang [51].

The gripper design inspired by the fin ray effect can achieve active deformation, which helps to simplify the trigger system and improve subtle manipulation capabilities [52]. With such a gripper, Jiahao Lin et al. [53] selected tomatoes and nectarines as fruit samples for experimental confirmation when determining the degree of fruit ripeness. The authors of this work claim that advances in soft robotics have demonstrated the potential for compliant deformation, safe manipulation, and skin sensing, providing many new solutions in various fields, including agriculture (Figure 6).

The Venus Flytrap, a plant that captures insects by trapping them between its two leaves, is an excellent grasper in nature [54]. It is known that several soft trap designs inspired by the Venus Flytrap were invented. Shi et al. [55] used IPMC actuators to create a miniature robot, taking inspiration from Venus Flytraps. This robot is able to achieve a tip displacement of 10 mm and can carry a payload as high as 36 mN. The dielectric elastomer (DEA) actuated soft gripper is able to open up its two-leaf structure when voltage is applied [15].

Pneumatic actuation of the gripping motion is most popular in a class of grippers designed for geometrically fixing objects. Such grippers often use fingers/jaws, which are the appendages of the gripper that actually make contact with the object. In the case of pneumatic actuation, the fingers are either attached to the mechanism or are an integral part of the mechanism [56].

Finger motion is based on soft chambers with pre-set deformations for targeted applications; the grasp of various objects is realized using positive pressure. A few configurations of two- and three-finger soft grippers based on pneumatic telescopic soft actuators were proposed by L. Gerez [57].

The detailed modeling and analysis of a soft robotic pneumatic actuated finger-like gripper (Figure 7) for agriculture, medical, and industry purposes mounted on an articulated robot were proposed in [58]. S. Dilibal et al. [58] presented the gripper, which consists of a three-finger structure designed as a mono-block to reduce the manufacturing complexity of the gripper.

A pneumatic soft gripper design based on the anatomy of the human finger was proposed by G. Udupa [59]. B. Wang et al. [60] reported a 3D multi-segment soft pneumatic actuator (3D-SPA) with a strong shell and multi-directional independent bending whose design was inspired by the multi-segment biological finger structure. Experimental results showed that, at a pressure of 40 kPa, the single-segment SPA reaches a maximum bending angle of 175.2°. The push and lifting forces of this soft gripper achieved maximum values of 11 N and 4.08 N, respectively.

Honjung Li et al. [61] offer soft pneumatic endoskeleton-type grippers made by 3D printing. The action of the finger is provided by air chambers in the fingers. The rigid endoskeleton, like a wristwatch chain, transmits the gripping force. However, as the authors declare, this product does not withstand lateral loads due to the low stiffness of the soft material. Z. Zeng [62] proposed another interesting solution, a pneumatically actuated soft gripper based on the use of bi-stable structures. The main advantage of such a gripper is the ability to operate without constant pressure to achieve the reversible shape transition of the bi-stable structure. Developers rated this gripper as a prospective solution to reduce power consumption in vacuum environments and underwater.

Shih et al. [14] present sensor skins (Figure 8) integrated into the fingers of soft grippers designed to operate with unknown delicate objects with feedback ability by modeling objects through touch. Each finger is prepared using silicone elastomer to realize complex movement. Flexible sensor skins on fingers are applied to measure contact and deformation for a better understanding of the environment where the gripper is working and to increase the safety of human–robot interaction.

Vacuum grippers can be classified according to suction cup material, type, the technologies used to create the vacuum, and the object materials’ geometry and surface roughness. E. Brown et al. [32], Courchesne [63], J. Y. Lee [64], and Amend [65] demonstrate (Figure 9) a different approach to the universal soft gripper working on the geometrical fixation principle. In this approach, individual fingers are replaced by a mass of granular material that flows around and conforms to the shape of the object when it is pressed during the grab. When a vacuum is applied to the chamber, the granular material shrinks and hardens to grab and hold the object without sensory feedback.

The relevant operating principle of such a gripper is the ability of granular materials to transition from a non-cohesive, deformable state to a frozen state with solid stiffness when friction, suction, and blocking contribute to the holding force. Due to the available large area of contact, such grippers provide higher stability in comparison to claw-like grippers and are adaptable to curved surfaces [66]. Furthermore, the vacuum pressure in this device enables friction control between the layers inside the construction of the gripper, and this effect is called a layer-jamming phenomenon [66]. The orientation of the object to be grasped plays an important role in the grasping procedure, especially in agricultural applications, such as broccoli processing, where additional object reorientation could be required. This action can be performed using a pneumatic variable-curvature soft gripper for object orientation mainly consisting of three variable-curvature soft fiber-reinforced bending actuators [66].

N. Hou proposed an origami gripper with pressure and elongation control that can capture a drone that lands vertically on the gripper [67]. A combination of the typical finger-based soft gripper controlled using a layer-jamming phenomenon is proposed by Y. Li et al. [68]. He proposed a method to bond a silicone rubber soft actuator and pack particles inside to form an integral gripping finger. By inflating air pressure control, the particles jamming inside control the stiffness of the finger.

A water-hydraulic flexible gripper with three flexible actuators combined with soft skin and an inner skeleton to resist water column pressure for underwater manipulation with deformable objects is proposed by H. Ji et al. [69]. The authors provide a mathematical model describing the relationship between water pressure and gear deformation. The simulation analyzes the effect of different materials, structure sizes, and inlet pressure. A Kriging model is used to fit the simulation results, and a multi-island genetic algorithm is used to obtain the optimal solution for the flexible drive structure.

K. C. Galloway et al. [70] proposed boa-type fiber-reinforced actuators and bellows-type actuators—which can gently conform around objects with the control input of a single hydraulic line. The result of this work, according to the authors, is the first application of soft robotics in the deep sea for non-destructive sampling of benthic fauna.

Magnetic actuation with applied principles of magnetism is often used for grabbing and manipulating ferrous workpieces [56]. Nevertheless, magnetic actuation is quite a popular solution for soft grippers. Permanent magnets have the advantage of not requiring an external power source to operate the magnet since operation using an external power supply sometimes could be challenging. Several methods have been developed to integrate cooling liquids into electric actuators to improve magnetic-type grippers [71]. Piezoelectric elastomers and high-entropy magnetoelectric alloy composites, proposed by X. Hu [72] and applied for self-sensing flexible grippers, demonstrate a new quality level of intelligent industrial equipment. The magnetic field in single-nanowire grippers working due to Lorentz forces with vibration tools to overcome van der Waals forces enables grasping with large amplitude of a microscale object, as proposed by J. Yan [73]. The release of small objects is an important stage of grasping, and then the release of small objects is realized by vibration [74].

Nishimura [75] proposed a gripper with fingers manufactured by 3D printing, with multiple different surfaces, resulting in nine available grasping configurations, with a single motor working on a self-motion switching mechanism utilizing permanent magnets. The gripper has a millimeter-scale grasping range for objects of different shapes and sizes, and it can also grip objects with parallel and enveloping motion.

Sun et al. [76] proposed a gripper (Figure 10) with an underactuated seesaw flexure mechanism with a pair of multifunctional jaws for enveloping and parallel grasping.

This gripper was tested by performing dynamic and static simulations and practical tests. Some hysteresis was noticed and simultaneously validated its ability to catch small irregular-shaped objects.

Electromagnets and magnets used as robot arm links or gripper actuators can be divided into gears with magnetorheological fluids and actuators using magnets and electromagnets for manipulating objects [71]. A liquid (oil)-cooled laminar electromagnet (LCLE) employed as a soft gripper for grasping heavy objects was presented by M. Rodriguez [71]. In his paper, he proposed a gripper that is controlled by changing its geometrical parameters by heating/cooling the oil used to cool down the magnetic actuator.

The grippers operating on a geometric bonding or embracing principle are efficient and can operate with many types of objects. Nevertheless, there are some problematic issues with geometric clamping. In the case of soft or fragile structures, there are high requirements for conforming grippers and object geometry. In addition, this restricts the quality of the gripper positioning in relation to the object-fixing surface.

On the other hand, geometric embracing can operate quickly and have a fast and reliable clamping procedure with minimal clamping force control; in many cases, it can be just enough geometrical support. This opens the possibility of using pneumatics as actuating power and simplifies entire design and operation matters.

4.2. Grippers Working with Friction Force

Friction force in gripper manipulation is evaluated ambiguously, although the main rule is that friction forces corresponding to the object fixation must be higher than inertia forces that could appear during the manipulation procedure. In some publications, authors have proposed a method to eliminate object slip by pre-calculating the required grip force and applying tactile sensors to estimate friction coefficients to determine the minimum squeeze force for a successful grasp [77] in the case of finger-based grippers that squeeze the object but do not lock it geometrically.

The most typical example of grippers that use friction force to hold the object is vacuum grippers with suction cups. In this case, the atmospheric pressure creates a force pushing the object against the suction cup surface. This gripper type is very popular in industrial robot applications, and the prior literature focuses on using vision-based planners (Figure 11) to improve grasping success in these tasks [78]. Haptic exploration for improving the suction cup grip is proposed to avoid failure of gripping when visual planning is complicated [38]. The gripper with two fingers of hemispherical shape and an RGB-D camera (Intel Realsense D345i, Intel, Santa Clara, CA, USA), proposed by F. Visentin [79], employs geometrical fixation and friction forces for the grabbing of very fragile fruits or berries. The weaving gripper mechanism (made from woven textile strip loops) was manufactured in different sizes for large and small object grasping, as proposed by G. Kang [80]. The origami-inspired gripper, manufactured as a bare leaf with ribbons inside, is a universal one that suits any object, as proposed by Y. Hong [81].

A study by Yuseung Jo et al. [78] presents (Figure 12) a suction-cup-based soft robotic gripper for cucumber harvesting that adjusts its shape and surface parameters to respond to the surface and shape characteristics of cucumbers. The shape of the gripper is optimized for adsorption to the large curvature of the cucumber, increasing the effective radius associated with the grip force. The suction cup surface is also modified to maintain adsorption on uneven cucumber surfaces. The suction cup test verifies the validity of these critical parameters and shows that the proposed copper provides robust adsorption, increasing the adsorption success rate and effective radius. The mechanism of this gripper, represented in Figure 12a, mimics a human harvesting method involving three subtasks, grasping, traction, and cutting, and the end-effector modules for cutting and grasping used to replicate the human harvesting method are represented in Figure 12b.

Park et al. [71] experimentally applied a developed vacuum gripper for different types of fruit and claimed successful harvesting using a gripper with a cutting module and transporting module, with a good relationship between the forces applied to the fruit due to suction.

S. Song et al. [83] proposed a soft robotic gripper where suction is generated in a self-sealing, fin-flat elastic membrane, which improves gripping. This gripper can independently adjust the size of the effective suction area depending on the applied load. The 3D printing here is evaluated as a convenient tool when emerging new material fabrication opens opportunities for the improvement of grippers. In addition, magnetic stimulation as a shape-changing mechanism allows printed actuators to control various functions of soft robots. The authors of this paper declared some weak points of this mechanism. Although permeability will require periodic charging of the negative pressure difference inside the gripper body, it has a low influence on performance.

An underactuated and vacuum grasping combination is proposed by Maggi et al. [84]. In this work, the gripper (Figure 13) is equipped with suction cups and manufactured to harvest pumpkins, apples, and other similar-sized plants.

H. Li et al.’s [85] work proposes a versatile method to equip the tips of soft bending actuators with bioinspired force amplifiers to greatly increase the payload of multi-finger soft grippers, where friction force is used to overcome gravity. The experiments of this work show that the proposed soft gripper composed of four bending soft pneumatic actuators has a maximum gripping force of ~1960 N, which is more than 17 times the current level, and can lift the human body. Also, this scientific group claims that the locking soft gripper also has the following limitations: soft bending gears must be arranged symmetrically in pairs because an odd number of soft gears cannot be equipped with all force amplifiers; contact-type force-amplifier fingers require the locking soft gripper to cover the object to be grasped, and the force-amplifier fingers may fail to lock when disturbed by external forces; and the locking soft gripper is inferior in dexterity to the classic multi-finger soft grippers.

As a result of normal attraction, the contact interface also produces a high shear friction force, which is useful for soft grippers when grasping and holding objects. There are two typical adhesion technologies, namely gecko-inspired adhesion and electrostatic adhesion. A gecko-inspired adhesive improves the grip and friction force of all types of grippers. Gecko-inspired glue was motivated by the micro- and nanostructures found on the animals’ toes. Nanostructures are directional—they bond strongly when sheared in a certain direction but weakly when sheared in other directions or under no shear load [86]. Electro-adhesion on supramolecular ion-tronics and microfabrication by a gel–elastomer interface enables ionotronic adhesive soft grippers to miniaturize grippers [87]. D. Gao proposed ribbon-like gripper fingers for microscale objects. M. Kang reviewed adhesives as materials suitable for involvement in soft grippers [88].

J. P. Roberge [77] analyzed a use case with “tactile exploration” and explained how friction generally works at the molecular level. Friction forces consist of two components: molecular attraction and molecular collision. The area of contact on the molecular scale is the first adhesion former, and, later, the load-controlled component is the resulting operation of frictional force [89]. Based on intermolecular forces, J. P. Roberge [78] proposed a gecko-inspired gripper that realizes the dry adhesion principle.

Gecko-inspired adhesives that could be applied on the gripper’s active surface have been mentioned in about four hundred articles over the last two years. One of the most popular trends is the application of high-aspect-ratio micropillar arrays. Biomimetic dry adhesives, superhydrophobic and tunable wetting surfaces with employed capillary forces, are also waiting for broader applications in soft grippers [90,91].

Inspired by natural dry adhesives, micropatterned functional surfaces found their way into automated gripping solutions as novel end-effectors for robotic pick-and-place applications. The functional surfaces consist of fibrillar structures, technically called pillars, ordered in patterns and commonly fabricated from soft elastic polymers. Analysis showed that differential stretching of fibrils was reduced, effectively lowering the misalignment angle and the resulting load concentration. These results are essential for applications of fibrillar microstructures under normal loading conditions without precise control of alignment, for example, in pick-and-place gripping systems [92].

Design strategies for utilizing the unique properties of gecko-inspired adhesives are presented by Glick et al. [93]. By modeling liquid elastomer actuators as joints with associated joint torques, we created an actuator that takes advantage of the unique properties of gecko-inspired adhesives. Experiments have shown that grippers are stronger at lower pressures than non-gecko drives, so, in most cases, the gripper can operate faster and use less power.

However, using these gripping systems still requires precise manual adjustment of the gripping parameters. To address this limitation, a coil-based sensor was designed to enable automatic detection of the attachment process. Herter et al. [94] presented a sensor that consists of three sensing layers (a transmitting coil, a conductive film, and a compliant). The sensor system allows precise detection of the misalignment angle and fast estimation of the qualitative direction of misalignment with minimal compression. Micropatterned dry adhesive systems are a promising alternative to conventional handling solutions.

Electrostatic adhesion is a type of electrostatic attraction that occurs between oppositely charged particles. Because the interface must be charged, this method usually requires a large external electric field. Electrostatic adhesion has been shown to be effective for manipulating objects with smooth and rough surfaces. Soft robotic grippers based on electrostatic adhesion have been reported in [95]. Sh. An et al. [96] propose a soft gripper with variable structure and electrostatic adsorption. The gripper consists of three fingers and can be adapted to grip objects of various sizes by dynamically modifying the opening angle of the fingers. The gripping of small flat objects is enhanced by installing electro-adhesion (EA) films on the corresponding surfaces of the fingers of a soft pneumatically actuated gripper.

The soft gripper proposed by V. Cacucciolo [97] involves (Figure 14) principles of electro-adhesion and dielectric elastomer actuators for the delicate manipulation of fragile objects. Here are two different electro-adhesion patterns at the fingertips that we used in this study. This gripper can generate actuation with a monolithic structure with a single electrode for a wide variety of objects. The authors point out that there are two widely used actuation technologies for this type of soft gripper: fluidic and electromechanical. Fluidic soft actuators can generate large forces and can be manufactured relatively simply, while electrically powered soft actuators can be implemented as simpler and lighter systems using compact actuator electronics.

Electro-adhesion soft grippers have advantages in grasping, including silent operation and low energy consumption. In [98], V. Cacucciolo hypothesized that the main factor for the grasping force of electro-adhesion grippers is the peeling angle between the electro-adhesive surface on the inextensible and elastic tapes and the object. Research on crack propagation without friction showed that, where pulling force uses the energy required to create a new crack, gripping force is lower.

M. Mastrangelo proposes another interesting electro-adhesion soft gripper (Figure 15) [99].

This gripper involves fingers wrapping around the object by zipping an electro-adhesion finger on a cylindrical object. It works as follows: electrostatic forces are created by the voltage difference between the electrodes, which attract the finger to the curved base and thus adjust the wrapping angle until the finger is completely fixed on the object.

A 3D-printed flexible gripper composed of a flexible electro-adhesion (EA) pad (Figure 16), a bioinspired spiral spring, and an EA fixture to address the limitations of current flexible-electrode and soft gripper fabrication methods is presented by Ch. Xiang et al. [92].

Figure 16 presents novel 3D-printed, simple-to-implement, cost-effective, and ready-to-use flexible grippers composed of flexible electro-adhesion and a flexible, adaptive gripper. Two types of conductive and non-conductive PLA materials were used to fabricate the flexible comb electrodes, 3D-printed TPU materials were used to fabricate the customizable, flexible EA mount, and, finally, the fully 3D-printed flexible mounts were fabricated.

Adhesive-type soft grippers for manipulations in the underwater environment, where the device uses a peristaltic pump and contains a “DragonSkin” adhesion layer, designed for delicate underwater living, are presented in [100] by N. Sinatra. This study represents soft robotic grippers capable of grasping delicate examples of gelatinous tissues. Thus, in this work, nanofiber-reinforced soft actuators apply low enough contact pressure to ensure minimal damage to typical jellyfish species. The development of this technology has the potential to improve various in situ characterization methods used to study the ecological and genetic characteristics of deep-sea organisms.

The chameleon’s tongue, as is mentioned by J. Qu [101], is known for its ability to expand when hunting or to catch object.

Grippers with friction as the main object-fixing factor in the clamp fingers or on the surface generate pressing force in a direction perpendicular to the friction surfaces. The success of the clamping solution depends on what opposes the motion between the two surfaces; the adhesion is the most important force for the best gripping result when friction force is employed. The success of the clamping solution lies in the proper implementation of the coexisting effects, like adhesion, van der Waals forces, hydrogen bonds, electrostatic forces, and, in some cases, chemical bonding. Although van der Waals forces operate only in the nanometer range, resulting in the adhesion between two materials, their relative values are significant.

Hence, all friction-force-based grippers’ working quality depends on the material properties. Polymers, plastic, and resins are usually used. Thus, the development of new materials with high friction coefficients or other friction-developing features remains important for friction-force-based grippers.

Quality of gripping—reliable clamping and saving of the object’s shape, including the shape of fragile items, soft items, or even biological objects with thin structures—remains an important issue in grippers with friction as a fixing force.

4.3. Grippers for Complex Gripping Method

Most soft grippers operate based on the application of different methods, materials, and forces and often involve combinations of geometrical bonding and friction force to hold the object reliably. Some gripper systems are inspired by nature and act as neuro-prosthesis devices. G. Gu et al. [102] proposed a device with grippers with six active degrees of freedom under pneumatic actuation to control movement of a prosthetic hand with integrated fingers, which enabled tactile feedback so that individuals could regain primitive touch sensation, and real-time closed-loop control. Abeach et al. [103] presented a classical three-finger dexterous gripper that uses two designs of McKibben’s muscles [104], or the extensor muscles with the elongation function. When two types of muscles are arranged to act antagonistically, pressure increases in all pneumatic muscles and increases the stiffness of the system without changing the position of the fingers.

The new-generation of soft grippers, as was introduced by S. Jain [105], is enabled to reconfigure the workplace of a soft gripper by using cylindrical fingers for convex objects, broadening fingers for scooping tasks, and extending fingertips for thin and small loads (Figure 17).

A slightly similar approach with a three-dimensional 3D-printed multipurpose pneumatic soft gripper with fingers and suction cups that operate separately or simultaneously was proposed by Ch. Tawk [106], for manipulating objects with different weights, sizes, shapes, textures, and stiffnesses.

A multi-joint gripper similar to the human hand was proposed by K. Ham [107] and presents a soft gripper with variable stiffness that can be used to control stiffness during tendon traction. The soft grip has three variable-stiffness structures that act like fingers, and the stiffness can be controlled using two motors by winding the tendons. Without monitoring contact forces, such a gripper cannot be applied to handling objects with an unknown degree of fragility. Although the correct gripping force or gripper opening for each target can be found by trial and error, it requires an expensive force/torque sensor or a precise gripper position controller.

Control of gripping force remains an actual issue for many researchers. S-J. Huang et al. [108] proposed a smart gripper designed with an embedded distributed control structure to avoid the uncertainty of object mass and soft/hard properties. A communication signal here is specified to integrate the control cores of the robot arm and gripper to perform the functions of robot position control and gripper force control in sequence.

The alternative for accurate gripping force control could be spiral grippers inspired by nature. Instead of grasping things with hands or tentacles, climbing plants weave around neighbor objects, plants, or a pole to raise their organism parts high enough off the ground. This grip is based on the stems of twisting plants, which find support by rotating in large diameters around a certain axis (Figure 18). The soft plant-inspired gripper idea looks simple and requires only one channel pneumatic controller. It should be mentioned that fiber-optic sensors, in some cases, are built into these grippers. They are a high-birefringence (HiBi) fiber in a Sagnac loop configuration that allows the detection of twisting angle, object size, and the occurrence of disturbances [109].

The gripper (Figure 18) proposed by M Yang et al. [109] was produced from silicon rubber using a 3D-printed mold. The delicate structure of such grippers is often the limitation of their life cycle, and their lifetime is relatively short. Therefore, soft grippers, especially bioinspired soft grippers, create a need for self-healing structures and materials. For this reason, soft grippers are often designed using materials that can restore the initial structure through chemical, adhesive, and other effects.

Self-repairing soft grippers are resistant to mechanical damage or aging wear and ensure long-term functionality and energy efficiency. Furthermore, bioinspired soft grippers can incorporate energy-saving modes, passive actuation, and energy-recovery systems, which is crucial in prolonged biomedical procedures involving grasping or other manipulation needs [110].

Another interesting approach is bionic cone-shaped octopus-like arms. Such soft grippers operate using a combination of geometrical bonding and friction forces. While grasping, their fingers bend, bonding the object geometrically, and, at the same time, the suction cups placed on the finger’s inner surface produce additional force, securing the object.

Prototypes of such grippers were demonstrated by B. Fang et al. [111], Xie et al. [112], and M. Wu et al. [113]. Experimental results demonstrate the capabilities of dexterous manipulations and capability for complex tasks, such as ocean litter collection or archeological exploration.

In another report, M. Wu [114] presented a soft gripper (Figure 19) inspired by the highly developed grasping ability of the glowing octopus (Stauroteuthis syrtensis), enabled by an umbrella-shaped dorsal and ventral membrane between each arm and consisting of a 3D-printed linkage mechanism used to actuate a modular silicone mold. With the help of a 3D-printed joint mechanism, the soft gripper can grip objects of various shapes and dimensions, including flat objects, objects beyond the grasping range, irregular objects, scattered objects, and moving objects.

Another example of bioinspired soft grippers is those based on the sea anemone’s predatory behavior, mainly guaranteed using membranes, thanks to which the grasping action is carried out (Figure 20). The basic action is to pull the outer skin from the inner skin through the conduction of strain and rotate inward to suck in the target. By mimicking this predatory process [115,116], the idea of a simple object grasping the structure of the sea anemone is applied to the soft gripper. The centripetal manner of gripping by an enveloping movement of the pins is employed with a passive membrane disruption procedure [115]. The proposed gripper design is not a direct copy of the sea anemone body; it achieves multifunctionality and multimodal gripping due to a flexible gripping process and adaptive deformation technique. The sea-anemone-style actuator composed of magnetic-driven “flowers” under an external magnetic field can mimic animal behavior, which is shown in the paper of X. Wang [116]. The quality of these grippers’ gripping force is enhanced using membrane-shaped active surfaces, and therefore they are suitable for the intellectual adjustment to the object shape.

In addition to the membrane technology, Zhang et al. [117] proposed a bionic webbed-foot soft gripper with variable stiffness with an integrated pneumatic network with stretch and bend control and a negative pressure analyzing system. In this design, the gripping force is increased by mounting extra “skin” on the gripper.

As we can see from Figure 20, the combination of the geometrical gripping methods compared with the friction one resulted in a complex soft gripper, inspired by animals walking on wet surfaces, where the wet surfaces have weak acting forces, such as molecular interaction forces, surface tension, capillarity forces, and internal material structural stresses.

Implementing complex gripping methods opens a broad range of objects for safe and intelligent gripping. Mimicking and biologically inspired methods are widely used here, and the number of bioinspired technical solutions is not limited by existing ones. The development of these methods will hopefully foster biological research and enrich the design library of robotic end-of-arm devices in general.

5. Discussion

Thus, after partially introducing the purpose of the grippers, directions of activity, and emerging difficulties, the classification of grippers can be represented more conveniently. Most of today’s grippers can be divided into two main categories based on how their basic movements are achieved. The first category is grippers, the movement of which is performed in kinematic pairs, where structural deformation is not decisive. This category is applied to clamps with several rigid links and joints moving or rotating the pair of symmetric clamps. This type of gripper, as usual, has only one degree of freedom, a simple structure, simple operation, and good reliability. The weak feature of these grippers is a problem with adaptation to different shapes. The second category consists of soft grippers, where the deformation of structural parts results in the execution of movement. Theoretically, soft grippers can have an infinite degree of freedom for rigid joints and connections when interacting with objects, result in deformations, and enable the adaptation of a large assortment of objects with various shapes [27]. Overall, soft robotics is the most promising robotic system development. Soft robotic systems are considered safe due to their compatibility and provide great solution flexibility and option customization in terms of design and material application abilities [118]. Thus, the obvious advantage of soft grippers led to greater interest in developing them in contrast to hard grippers designed for a specific object [39].

According to the desired function, grippers can be divided into the following groups:

• Soft grippers for fragile objects [119];
• Soft grippers for objects of indefinite shape [120,121];
• Smart grippers with macro sensors [122];
• Smart grippers with advanced customization [123,124].

According to their types of movement and a classification model inspired by biological “grippers”, soft grippers are classified into three types:

• Non-continuum bending-type grippers [125];
• Continuum bending-type grippers [28,126,127];
• Continuum twisting-type grippers [128].

Thus, attempts to systematize grippers across the board use a variety of approaches and evaluation cuts.

A detailed and up-to-date analysis of each gripper type is provided. In addition, this review reviews the various stiffness-controlled strategies developed in recent years.

In general, grippers can be classified according to the quality of the object to be gripped, the size of the object, and the gripper’s operating system. A brief review of the literature analyzed under this framework is presented in

Table 1. Classification of robotic grippers for various object properties.

Type of the Object	Size of the Object	Gripper Type	References
Strong, unsensitive	micrometric	mechanic	[129]
		electric	[122]
	10 … 100 mm	mechanic	[37,75,77,108,130]
		hydraulic	[131,132]
		vacuum	[133]
		pneumatic pressure	[28,29,30,31,103,104,106,126,127,128]
	>100 mm	mechanic	[23,134,135,136]
		hydraulic	[137,138,139]
		vacuum	[140,141,142]
		magnetic	[71,143],
		Coanda effect gripping	[144]
		adhesive	[145,146]
Fragile	micrometric	mechanic	[76,147,148]
		hydraulic	[149,150,151]
		vacuum	[152]
		pneumatic pressure	[153]
		adhesive	[154,155]
	10 … 100 mm	mechanic	[84]
		hydraulic	[156]
		vacuum	[157,158]
		pneumatic pressure	[159]
		electric	[160,161,162,163]
		magnetic	[164,165,166]
		biomimetic	[167]
	>100 mm	vacuum	[168,169]
		pneumatic pressure	[170]
		hydraulic	[171]
Soft and sensitive	micrometric	magnetic	[172,173,174]
		pneumatic pressure	[175]
		adhesive	[176]
		hydraulic	[28,177,178]
	10 … 100 mm	magnetic	[179,180]
		hydraulic	[69,70,86,100,181,182]
		vacuum	[32,38,78,82,111,112,113,114,115,117,183,184],
		pneumatic pressure	[22,26,29,30,31,54,58,59,60,61,62,96,105,109,154]
		adhesive	[91,92,97,98,99,123,124]
		particle jamming, pressure	[68]
	>100 mm	mechanic	[21,27,30,53]
		pneumatic pressure	[185]
Special case		vacuum, for climbing	[186]
		capillary forces	[187]
		hydraulic, deep sea	[188]
		electric, nanoscale	[73]

below.

Evaluating the collected information represented in the table, we can draw several conclusions. According to the current needs, the grippers are relevant for gripping both fragile and heavy objects. Therefore, grippers of both types are currently needed and have a future. Biologically inspired grippers have several advantages; there is no need to fantasize because nature has tested the most diverse strategies and left the most adapted to work. Simply copying biological systems is not enough; a very detailed study of those systems is required, as well as an understanding of what produces the desired result and how, all the more so because the structures in nature are based on ensuring operation at the molecular level. There are quite a number of scientific publications revealing how the chemical structure of the material and its modification influence the quality of the gripper’s performance.

6. Conclusions

After systematizing the collected information from scientific publications, most grippers cannot cover many functions and areas of application. Currently, more attention is being paid to grippers for the assembly of smaller objects, as new technologies, new materials, and combinations of these join together to create more sensitive and versatile grippers.

Grippers and manipulators play an essential role in the process of interaction between creatures and soft robots and their environment. Considering various industrial and healthcare requests, a new generation of grippers has been developed with the advantages of friendly operation, high compliance, and compatibility. Soft grippers can perform gripping and handling tasks through applying adhesion or friction as a new bridge between robots and the objects to be manipulated. However, most of the existing soft grippers are too limited by their soft materials and soft actuators, which are affected by high temperature, high pressure, cold, and corrosive environments, to meet the needs of research in extreme environments such as space exploration, ocean depths, and polar regions. Soft grippers tend to age and fail quickly when working in extreme environments without any protection. Thus, learning from creatures that live in extreme environments may be one way to overcome these problems. The soft structures are created by applying various stimuli such as pH, temperature, magnetic field, and many combinations of soft materials. The 3D printing of stimuli allows for a variety of shape changes such as bending, twisting, folding, swelling, rolling, shrinking, origami, or movement. A wide variety of soft magnetic structures can be produced by embedding soft or hard magnetic particles into soft materials, leading to magnetically active soft materials.

Hence, the smaller the object to be grasped is, the more influence intermolecular forces employed in the grasping process have. The materials applied for the grasping of large objects seem to escape the influence of weak material forces, but even large structures present problems related to the specifics of the material structure and possible hidden internal defects, such as natural fatigue of material, which is practically a consequence of intermolecular interactions.

Each field of gripper application has specific requirements for operating the gripper safely. For example, grippers manufactured for medical purposes can have a body-friendliness requirement, with limitation of a certain material, and specific knowledge requirements may arise. When the object to be grasped is an agricultural product, arising challenges may be eliminated by practical training before reaching the real agriculture field. The situation where the gripping object is fragile, unknown, and in an underwater area is more complicated and authors of suitable devices for such a situation present as usual detailed report of practical tests.

New gripper control technologies and new software exclusively expand the application possibilities and efficiency when the gripping process is well designed and unchallenged. The quality of future grippers can be improved for products developed using new chemical materials, polymers, metal alloys, and joint development of gripper fingers. There is a separate developing area for nano-grippers, which, like nanorobots, operate at the molecular level and differ from conventional grippers in terms of implementation principles, and require specific knowledge, equipment, and fields of applications. Application of 3D printers and drone-mounted grippers allows rapid development and improvement of gripping capabilities.

Therefore, gripper developers must constantly follow innovations in material science, programming, design, and control.

With a compatible mechanism for the gripper, the gripping force would greatly depend on the flexibility of the material; therefore, by choosing the right material and the right design, a soft gripper can be used with a compatible mechanism to achieve the target gripping force. From a basic structural perspective, joints and joint changes will alter overall function and constraints. Simple structures are more adaptable. However, there is no doubt that more complex and flexible soft gripper structures are a future research direction. Four-dimensional (4D) printing, also known as the AM of smart materials, is gaining popularity among the scientific community as it has great potential to produce soft structures such as soft robots, actuators, and grippers.