Research on various types of soft robots has been conducted recently. Soft robots with flexible mechanisms are expected to contribute significantly to the development of robotics, not only in industrial applications but also in various other applications, such as those involving robot-human interactions. Flexible mechanisms cannot necessarily be realized using flexible materials alone; they can also be realized by adopting new approaches to conventional mechanisms. It is also desirable for sensors and other components to be flexible so that they do not reduce the flexibility of the mechanism. The development of these mechanisms and elemental technologies will lead to further development of soft robots.

This special issue will feature papers on the structures and applications of soft mechanisms and elements as well as on design and analysis methods and control techniques.

Flexible/stretchable sensors comprising soft structures that do not interfere with the softness of the bodies of soft robots are essential for achieving soft robots with superior operational performance. These sensors are expected to be applicable as sensing skin for humanoid robots and as interfaces for human-robot interactions. Herein, we refer to flexible/stretchable sensors investigated for soft robotics applications as “soft sensors” and review recent research trends. Specifically, we discuss optical, resistive, capacitive, and inductive soft sensors with emphasis on their materials and structures.

Recently, soft actuators have gained considerable attention owing to their flexibility and high output-to-weight ratios. The McKibben pneumatic actuator (MPA), a type of soft artificial muscle, is an actuator that generates force by inflating a rubber tube with compressed air. Conventional MPAs, such as linear actuators, generate force along straight lines; hence, achieving complex movements, such as bending using a single muscle, can be challenging. In this study, we enabled bending movements in MPA by applying an elastic adhesive coating to MPA. Experimental results demonstrated that the coated MPA successfully performed bending movements. Furthermore, we confirmed that the curvature and fiber angles of the coated and uncoated surfaces changed with applied pressure, thereby indicating that the adhesive can be used to control the fiber angles and achieve the desired curvature.

Soft robots have promising applications in healthcare, manufacturing, and disaster relief. Inspired by biological models, their superior flexibility and environmental adaptability have increased their use in soft grippers. However, soft grippers often struggle with precise force control and high energy consumption during operation. One solution to these challenges is the utilization of electrostatic adhesion technology. This technology uses electrostatic forces generated by applying a high voltage between electrodes to adhere to objects, allowing for energy-efficient and delicate manipulation. We propose a novel paper-based origami electroadhesive pad (p-OEP) utilizing paper mechatronics. Thin-film comb electrodes were formed through inkjet printing, and self-folding technology was employed to create a structure that adapts to the curvature of the target object. This increases the adhesive area between the p-OEP and object, achieving high strength. This innovative approach expands the application range of soft grippers and provides a new gripping solution that is low-cost and environmentally friendly.

With aging society, pneumatic soft actuators with human friendly features such as welfare support and virtual reality have been developed. In a previous study, a flexible pneumatic spherical actuator (FPSA) that could allow individuals to exercise passively while holding on to the top end of the actuator during motion was proposed. The actuator’s ability to improve the range of joints after surgery was tested. The FPSA comprises three extension-type flexible pneumatic actuators (EFPAs), shaped like a rugby ball through restraint by PET sheets. This paper introduces a non-contact bending/extension sensor composed of a Hall sensor and a ring-shaped magnet, proposed and tested as an attitude sensor. Recognizing that the FPSA’s bending direction depends on the pressurized EFPA, a simple measurement system was also proposed and tested. This system measures the bending angle and extensional displacement by measuring the displacement of the restraint plate at the center bottom of the FPSA. To simplify the attitude control system, a design reducing the number of valves from six to four was developed and tested. Results confirmed that the FPSA, equipped with this control system, could trace the desired bending angle even when the bending directional angle changed. Additionally, an analytical model was proposed to evaluate the effectiveness of passive exercise enabled by the spherical actuator. This model calculates the FPSA’s shape for various input pressures to the three EFPAs in open-loop attitude control. Results showed that the calculated attitude closely matched the actual actuator’s behavior.

Nowadays, owing to declining birthrates and an aging population, patients and the elderly requiring rehabilitation are not getting enough physical activity. In addressing this issue, devices for rehabilitating them have been researched and developed. However, rehabilitation devices are almost exclusively used for patients who can get up, rather than those who are bedridden. In this study, we aim to develop a rehabilitation device that can provide passive exercise for bedridden patients. The ankle joint was selected as the target joint because the patients who have undergone surgery for cerebrovascular disease remain bedridden, and early recovery in the acute stage is highly desirable. We proposed and tested a fan-shaped pneumatic soft actuator (FPSA) that can expand and bend stably at angles when supply pressure is applied as an actuator for a rehabilitation device to encourage patient exercise. However, the previous FPSA’s movement deviates from the arch of the foot owing to increased supply pressure. In the ideal case, FPSA should push the arch of the foot in an arc motion. This study proposes and tests the FPSA that can operate a bending motion to provide passive exercise to the ankle joint using tensile springs and a winding mechanism powered by a servo motor. The proposed FPSA has a significant advantage of exhibiting no hysteresis in its pressure-displacement characteristics. The configuration and static analytical model of the improved FPSA are described.

The goal of this research was to design an artificial hand and bow force-adjusting system for playing the violin and to validate its effectiveness. The hand had five solid fingers without rotary joints. Instead, a chamber covered with a silicone rubber membrane was placed in each finger, except for the thumb. When the air pressure in the chamber is increased using an electric pneumatic regulator, the membrane inflates, which increases the force required to push the bow stick. We fabricated a hand and installed it on our 7-degrees-of-freedom manipulator. We conducted preliminary experiments to obtain the basic characteristics of the hand and built a bow force-adjusting system. We performed an experiment to confirm that the bow force could be adjusted by increasing or decreasing the air pressure in the index and little fingers. Thus, the bow force was successfully adjusted to the target using the artificial hand.

Recently, Japan has been witnessing an increase in the average age of agricultural workers and a decrease in the number of new entrants into farming, both of which are progressing year by year due to the country’s declining birthrate and aging population. As a result, expectations for substitution by robots and human-robot collaboration are rising. Therefore, we propose a robot arm built using straight-fiber-type pneumatic artificial muscle (SF-PAM) and a noncircular pulley. SF-PAM is sealed and has no sliding parts; thus, it has excellent dustproof and waterproof properties and is suitable for work on farms. However, due to its structure, the SF-PAM has a nonlinear relationship between the contraction force and the amount of contraction, and the output torque is insufficient near the limit of its range of motion. As a solution to this problem, a noncircular pulley is introduced to compensate for the output torque and expand the range of motion. Based on this, this study aims to realize fruit harvesting operation using a robot arm. In this paper, a two-degree-of-freedom robot arm was developed, and position control experiments were conducted to verify the tracking with the target value. As a result, the mechanical equilibrium model of the wire-pulley mechanism was found to be valid for this robot arm. However, issues were found due to the arrangement of the SF-PAM and the shape of the noncircular pulley.

This paper discusses model validation of hydraulic joints in the presence of unknown external forces using ready-made components. First, we review hydraulic cylinder dynamics with nonlinear pressure dynamics. Second, we design a controller that applies existing methods so that hydraulic joints have high backdrivability against unknown external forces. Third, we discuss the experimental and numerical responses of hydraulic joints in the presence of unknown external forces as the model validation. No custom-made cylinders, valves, or pipes are used in our experiments. Remarkably, the experimental and numerical responses were close to each other in the presence of unknown external forces.

This paper proposes a robot that combines a hyper-extension unit at the front with an earthworm-type robot at the rear, designed to inspect long, narrow, and complex pipelines. In conventional in-pipe inspection robots with simple designs, there was no mechanism for the robot to effectively exert traction and propulsion while optimizing speed according to the required force. Therefore, this paper describes the development of a robot that combines a hyper-extension unit at the front, which achieves an elongation rate of approximately 300% even at low expansion rates, with an earthworm-type robot at the rear. The developed robot can operate in two patterns: pattern 1 combines the movements of an earthworm and an inchworm, while pattern 2 resembles that of an inchworm. It can switch between speed and force modes by changing its motion pattern. Experimental comparison of the horizontal straight pipe movement speeds of pattern 1 and pattern 2 without and with load showed that pattern 2 was 3.32 times faster than pattern 1 without load and that pattern 1 was 1.51 times faster than pattern 2 with load, clarifying the importance of changing robot’s motion pattern according to the load. The vertical pipe movement speed without load was 3.04 times faster in pattern 2 than in pattern 1. The proposed robot traveled 1.43 times faster than the conventional robot equipped with a conventional spring mechanism when passing through the elbow pipe using pattern 1. However, it could not pass through the elbow pipe using pattern 2, showing the importance of changing the robot’s motion pattern when passing through bent pipes. These results indicate that the proposed robot has the potential to achieve more efficient inspection in long-distance, small-diameter, complex pipes by changing its motion pattern according to the situation in the inspection pipe.

One of the goals of soft robotics is to achieve spontaneous behaviors similar to real organisms. To gain insight into these behaviors, we examined the long (16-hour) spontaneous exploratory locomotion of Tegula nigerrima, an active foraging snail from an intertidal rocky shore. Specifically, we tested the critical brain hypothesis that the nervous system is inherently near a critical state that is self-organized to drive spontaneous animal behavior. The hypothesis was originally proposed for vertebrate species, but may also be applicable to other invertebrate species. We first investigated the power spectra of snail locomotion speed (N=39). The spectra exhibited 1/f^α fluctuations, which are a signature of self-organized criticality. The α value was estimated to be approximately 0.9. We further examined whether the spatial and temporal quantities showed multiple power laws and scaling relations, which are rigorous criteria for criticality. Although these criteria were satisfied over a limited range and provided limited evidence of self-organization, multiple power laws and scaling relations were satisfied overall. These results support the generalizability of the critical brain hypothesis.

The class Gastropods, which includes snails and sea slugs, inhabits a wide range of environments. Members of this class move by utilizing waves generated through muscular contractions in their soft body tissues (pedal wave). This characteristic is observed in both aquatic and terrestrial species and serves as a mechanism for adapting to various environments. In this study, we developed a soft robot that generates pedal wave using soft matter. This soft robot employs sensing based on machine learning, utilizing the soft material as a physical reservoir to leverage its environmental adaptability and soft exterior. The experiments investigated in this study were the relationship between locomotion performance and environments and the identification accuracy achieved through machine learning.

This paper proposes two position control methods for a double air chambers pneumatic artificial muscle (W-PAM) developed to improve the hysteresis characteristics of the rubber-less artificial muscle and confirms their characteristics and usefulness. The proposed control methods are the external pressure-regulated position feedback control and the internal pressure-regulated position feedback control. Of the two air chambers in the W-PAM, the air chamber regulated by the feedback control is different. The steady-state characteristic and frequency characteristics of these controls were compared, and it was experimentally confirmed that the adjusted position feedback control is superior in terms of position control performance. On the other hand, we confirmed that external pressure-regulated position feedback control can express passive stiffness, which is an important characteristic of pneumatic artificial muscles. These results indicate that the W-PAM is a useful actuator for human coexistence systems, because it can express both positional control performance and softness by using different control methods depending on the application.

McKibben pneumatic actuators (MPAs) are soft actuators that exert tension by inflating a rubber tube with compressed air. Although electropneumatic regulators can control air pressure, their cost and size limit their applications. This study employs a dynamic quantizer to control an MPA using a small solenoid valve that can only open or close, as opposed to an electropneumatic regulator. A dynamic quantizer is a type of quantizer that converts continuous signals into discrete signals. Our previous study confirmed that the tension or length control of MPA can be achieved using a dynamic quantizer. As MPA exerts force only in the direction of contraction, multiple MPAs must be combined when using them as robot actuators. This study demonstrates that control using a dynamic quantizer is feasible, even when multiple MPAs are employed. We focused on a pendulum driven by two MPAs to achieve angle-tracking control using a dynamic-quantizer-based control method. The results of numerical simulations and experimental tests confirm that the angle of the pendulum can be controlled using MPAs with a dynamic quantizer.

In this study, we introduce research aimed at developing a compact particle-excitation flow control valve that enables continuous control of air flow rate. We have been developing a compact control valve with continuous flow control for application to soft actuators. The control valve that we are developing uses particles as a valve plug, which does not require a valve plug positioning function and allows the control valve to be miniaturized. Furthermore, by controlling the movements of the particles on the orifice, the control flow rate can be made finer and the response can be improved. On the other hand, the past driving principle used the resonant mode of an oscillator utilizing a piezoelectric element (PZT) to control the motion of the particles. In this vibration mode, the size of the oscillator limits the miniaturization. Therefore, we propose a drive method that uses simple plate deflection vibration mode. This drive method eliminates the need for a large oscillator, and thus reduces the size of the control valve. To confirm the feasibility of the proposed drive method, we fabricated a prototype and evaluated its characteristics during flow control.

Actuators and robots capable of representing surfaces can take various forms, depending on the types of actuators used and their arrangements. In traditional robotic surfaces, the corners on the boundary in the undeformed state remain unchanged, indicating that the number and position of the boundary corners do not vary during deformation. This paper introduces a ring-shaped actuator with three types of bending elements combined with a soap film, demonstrating the existence of robotic surfaces in which the number of boundary corners can change through actuation. We also propose forward and inverse kinematics applicable to such robotic surfaces and present simulation results. These findings suggest that inverse kinematics can be achieved in soft robots constructed with prestressed silicone rubber or fabric, including membrane-like components, stretched over a frame.

Wearable power-assisted robots driven by artificial rubber muscles should be designed with a wide range of assistance capabilities to support various tasks while minimizing the sense of restraint when worn. Ideally, to realize a compact and lightweight power-assisted robot, the size of the artificial rubber muscle should be maintained while increasing the amount of movement. This study aims to develop an artificial rubber muscle that can achieve a wide range of assistance through rope-based force transmission. An artificial rubber muscle is proposed that can transmit force through a human-like pulling motion using a rope. This McKibben-type artificial rubber muscle is enhanced with a power transmission mechanism. The proposed artificial muscle has two layers: an inner layer that is deformed by air pressure to grasp the rope and an outer layer that pulls the rope by contraction. Each layer moves stepwise by supplying compressed air to realize the pulling motion of the rope. In this study, a prototype of the proposed artificial muscle was fabricated, and its operational principle was confirmed experimentally. Additionally, the force transmission performance was verified by comparing the contraction force of the artificial muscle with the tensile force of the rope.

In recent years, soft machine systems that employ soft actuators with high affinity and safety toward humans have attracted attention. The soft actuators developed thus far include those that can extend or bend by only supplying compressed air, or those that switch their motions by modifying their structures. However, the type of motion is limited and structural modifications are required to switch between different movements. In this study, we developed a soft actuator with multiple degrees of freedom that allowed switching between extension and bending without structural changes. This actuator comprised a structure in which flexible linear brakes (FLBs), which are negative-pressure-driven linear braking mechanisms, were arranged alongside a bellows-structured silicone rubber tube (bellows tube). The bellows tube extended when compressed air was supplied. When engaged, the FLBs provided fiber reinforcement against the bellows tube and bent the actuator. Thus, the actuator switched between extension and bending by engaging or disengaging the FLBs without structural changes. In this paper, we describe the structure of the actuator and the mechanical model used to arbitrarily change its bending direction using braking mechanisms.

A peristaltic mixing and conveyor that imitates the intestines of living organisms is used to mix powders and highly viscous fluids at low shear force in the manufacturing processes of foods, pharmaceuticals, paints, cosmetics, and other products. Although the device uses sequence control to mix and convey the contents, the actual intestine controls its motion autonomously according to the state of the contents. This research aims to establish an autonomous control method that can change between mixing and conveying according to the state of the mixture. In this paper, we propose a method to construct a learning model that is not affected by the content input conditions, in which samples with different total input amounts and mixing ratios are fed into the device, and sensor values acquired during the device operation are used as training data. The generalizability of the learning model was verified, and it was shown that mixing state estimation was possible up to 20 min from the start of mixing for mixtures with different input amounts and input order.

This study presents a novel method of fermentation promotion using a peristaltic pump for large-scale fermentation. This pump physically crushes, mixes, and diffuses fermentation substrates to mimic fermentation in the human gut. The proposed device effectively mixes fermentation substrates according to their properties and promotes fermentation. Preliminary studies were conducted on fermentation experiments using this device with fibrous and porous substrates.

This study addressed the problem of estimating the relative postures of two robots constrained by materials. The constrained robots were passively guided by human operators, and the relative postures were estimated using only encoder information from the wheels of each robot. This method relies solely on the encoder information and does not utilize additional sensors for relative posture estimation. A significant advantage of this method is cost-effectiveness and reduced maintenance requirements, as the robots do not need to be equipped with multiple types of sensors. The motivation for this research was to enable multiple omni-directional mobile robots to collaboratively transport long objects in confined spaces, such as those found at construction sites. Experimental results demonstrate that the relative distance between two robots can be estimated with an error of approximately 2%, and the relative posture with an error of approximately 2°. These results represent a notable improvement in the measurement accuracy compared to those obtained previously by the authors.

A two-joint robotic unit was developed to perform various tasks using a single unit. The unit is covered with a flexible plastic film and an insulating fluid, which makes it waterproof. A flexible film may limit the range of motion of a robotic unit. The robot packaging method, which is a method of packaging the internal components of a robot using the vacuum packaging technique, is used for plastic film covering. The effects of the film and the insulating fluid were experimentally clarified. Robot fingers, fish-like robots, and crawling robots were considered as application examples of the proposed robotic unit. These examples demonstrate that the robotic unit can work in contact with objects and in water.

This paper explores the innovative application of 3D printing technology in developing fins for underwater robots. The study delves into the design, fabrication, and optimization processes involved in creating fins using 3D printed technology, with the objective of improving the agility, performance, and overall capabilities of underwater robotics. The analyses and experimental results present the empirical findings of testing these 3D-printed fins.

This study analyzed upper torso rotation, which is crucial for improving pitch velocity and minimizing upper limb load during pitching. A simplified model was developed to facilitate the analysis which focused on torso torsion. In this model, the instantaneous torque transfer ratio and torso mobility were defined as key indices. We experimentally verified how these indices vary under three conditions: no torsion, static torsion, and dynamic torsion. The results demonstrate that torso mobility increases in conjunction with the instantaneous torque transfer ratio. Both static and dynamic torsion conditions resulted in higher instantaneous torque transfer ratios compared to the no-torsion condition. In static torsion, the maximum torso mobility was, on average, 26% greater than that observed under no torsion. In dynamic torsion, the maximum torso mobility was, on average, 69% greater than that observed under no torsion. These results indicate that torsion is effective in increasing both the instantaneous torque transfer ratio and torso mobility and that this effect is stronger in dynamic torsion. Therefore, torsion increased the angular velocity of the upper torso in response to torque input from the lower limb, potentially resulting in higher ball velocity and a reduction in upper limb strain. Additionally, the findings imply that the input required from the lower limbs to achieve a specified pitch velocity may be reduced. The simplified model and indices proposed in this study provide a foundation for designing exercise intervention techniques, evaluating athletic performance, and assessing injury risk related to torso rotation.

Robotic systems have been introduced to convenience stores to fulfill several menial tasks. One of those tasks is stocking, disposing, and arranging the product in the item rack. The general approach is to use a robotic arm to do pick-and-place tasks with the product. However, to solve this task, an appropriate algorithm is needed. In this study, we propose building an efficient algorithm for automating stocking, disposal, and arranging tasks with a robotic arm. In order to build it, a set of cost functions can be derived from the goal of stocking, disposal, and arranging processes in convenience stores. The derived cost function can then be used as a basis for creating a new algorithm that can drive the logical decisions of the system. A method to find temporary spaces within the item rack is also implemented in this algorithm. To find the most efficient form for the developed algorithm, several algorithms based on the cost function are built and tested on a simulated system of convenience store conditions. From the experiment, it is shown that, with the cost function created, using the greedy search algorithm performs the best compared to the other methods.

The vast ocean, which accounts for 70% of the Earth’s surface area, contains abundant minerals, energy, and biological resources, and surveys have been conducted in various ocean areas in recent years. It is extremely difficult for humans to conduct direct underwater surveys. Autonomous underwater vehicles (AUV) are expected to serve as platforms for surveying marine resources widely distributed throughout the vast ocean. However, owing to disturbances such as tidal currents, AUVs are unable to control there position well, making waypoint tracking difficult and preventing the observation of targeted observation points. In this research, to improve the robustness of AUVs in a tidal environment, we design a disturbance observer that estimates tidal currents as disturbances, and estimate tidal currents based on the results of actual diving surveys in actual sea areas using AUVs.

An actuator for achieving the same force in both directions, and a mechanism for supporting rotary motion with the actuator are proposed. By fabricating models and performing tests, the action of the actuator mechanism, and its application on an excavator was demonstrated. Simulations on the arm-moving task performed on a conventional excavator and an excavator with the actuator mechanism show an efficiency improvement with the use of the proposed actuator mechanism, particularly when the arm is driven at low speed.

This study proposes a novel respiration differentiation method using a wavelet transform and an autoencoder. Depth changes in a person’s chest were acquired from the depth images generated by the Kinect sensor and recorded as waveform data. The extracted waveform data were frequency-analyzed using a wavelet transform for the autoencoder learning. The autoencoder was trained exclusively on normal respiratory data. Respiratory data were identified based on the differences between the autoencoder’s inputs and outputs. To automatically differentiate respiration from other movements, a threshold was set using Hotelling’s theory. The proposed respiration differentiation method was experimentally evaluated to verify its high recognition rates, even for untrained individuals, using cross-validation. Respiration differentiation was also applied to the entire image to confirm that the model could accurately recognize chest movements resulting from respiration.