The mechanism and various control methods of the climber for the space elevator have been developed considerably in the range of operation on the ground. However, there are many problems when operating in the actual space environment. In this study, we developed a mini-climber experimental device with a loop-type tether for designing and developing a climber for the space environment. In the experiment, the mini-climber was stored in a vacuum chamber and compared its operating characteristics under atmospheric pressure and low vacuum. Under low vacuum, there is almost no atmosphere, and it was confirmed that the higher the rotation speed, the higher the operating efficiency. However, a significant temperature rise was confirmed because the heat generated by friction between the climber and the tether was not diffused. Various data simulating the space environment has become valuable data for climber design.

Exploration with rovers is one of main methods of lunar exploration. However, the rover may slip down or get stuck on the slope, whose angle is higher than 25 degrees, due to the loose ground on the moon. However, there is a moving method called push pull locomotion can effectively improve the climbing performance of the rover on loose slopes. This study validates the use of flexible wheels to suppress the slide of the PPL rover. Experimental results indicated that the rover using flexible wheel can reduce the slip.

Landing and moving for investigation on the lunar surface by rovers is an effective method. However, there is a possibility of a rover skidding causes it to deviate from its intended course. Therefore, it is necessary to obtain the driving state in real time. We focus on the strain generated in the rover's chassis. It is expected that the strain has a correlate closely with the driving state. The purpose of this study is to try estimating the slipping state of the rover in real time by analyzing the strain value measured with strain gauges mounted on the chassis. The measured data verified that the FFT analysis of the strain was different due to the slipping state of the wheel.

The supporting force that the leg of the rover is received is increased when vibration imparts to the ground. This phenomenon improves the running performance of the legged rovers. This study investigates the change in the supporting force by vibration in low atmospheric pressure. The atmospheric pressure of the Mars and Moon are lower than one of the earth. This environment affects the characteristics of ground sand because the air resistance is decreased. This study offers valuable information about using vibration in planetary exploration to improve the running performance of legged rovers. The experimental results show that the supporting force in low atmospheric pressure is almost the same as one in standard atmospheric pressure. It means that the condition of atmospheric pressure hardly affects the supporting force when imparting vibration.

In this paper, the diving and floating performance was evaluated for an omni-directional propulsion underwater robot (Cerchio). The Cerchio has a structure in which four fins are arranged at 90-degree intervals, and each fin has two-DOFs, the pitch axis and the yaw axis. The direction of diving and floating is controlled by the angle of the pitch axis, and the speed is controlled by the frequency and amplitude of the yaw axis. The performance was evaluated from the movement trajectory with the angle of the pitch axis as a parameter. From the results, the maximum horizontal deviation of ±10cm and the maximum rotation angle of 49deg were observed at present conditions. It was found that it is possible to control the propulsion direction and speed by adjusting the angle of the pitch axis. The maximum speeds of diving and floating were 0.080m/s and 0.098m/s, respectively.

The oceans occupy about 70% of the Earth ’s surface have vast space and mineral, energy, and biological resources. In recent years, AUV has been developed for resource surveys in various sea areas We are working on developing and implementing a robust control system to improve the AUV autonomy and to navigate safely. This paper reports about observation results by hovelling type AUV ”Tuna-Sand2” on seaweed bed habitat off Iwakuni on January 2023 and estimate currents from the dive data to design the observer to estimate currents and implement for AUV. The dive results show that there were Northwesterly current and AUV position is easily affected by current when change target direction.

Sperm whales prey on giant squid. However, the predation scene has not recorded on video. Bio-logging is useful for surveying animals that live in inaccessible area. In contrast, it is not easy to mount a camera to sperm whales. To tackle the problem, we have developed both an attaching system with a drone and a dropping-type adsorption logger. It is desirable that the logger can be attached not only to the horizontal body surface but also to the sloped body side surface. In this paper, we clear the effects of center of gravity for adsorption capability for the slope.

In recent years, the impact of the deterioration of the marine environment on ecosystems has become the problem. One of the causes of water quality deterioration is the hardening of sand and mud on the seafloor. Eelgrass beds, which provide a habitat for young fish, have been decreasing in number due to these effects. In this study, we focused on seabed tillage as a means of improving the hardening of sand and mud on the seafloor and creating an environment conducive to the growth of eelgrass. Currently, large-scale tillage using vessels is the mainstream, and it is not easy for a single organization to perform. This research aims to develop a small-scale subsea tillage robot that can be handled by an individual. Previous research until last year confirmed the rotation of the blades of the submarine tillage robot. However, a problem emerged in the mobility performance. In this report, we describe the process of manufacturing an underwater crawler and confirming its operation in order to improve its mobility.

We have developed a fish-shaped robot using snap-through buckling. In this paper, we focus on the intermittent and powerful movements of these robots and propose an application as a pseudo-live bait for fishing. We developed a small fish-shaped robot based on a scanned model of an actual horse mackerel. Swimming experiments were also conducted to simulate the swimming fishing. As a result, we succeeded in making the robot swim like a small fish, and showed that the robot can be used as a pseudo-live bait.

This paper presents the development of an eyeball integrated with a wide-angle lens camera, and a vision-sharing system for android avatars, aiming to provide both a comfortable sense of operation and a natural, human-like appearance for the avatars. The developed eyeball closely resembles a human eyeball and is capable of vertical and horizontal rotation. The vision-sharing system enables the operator to shift their perspective in a manner similar to how a human would in a real environment via the avatar, as the avatar’s gaze direction aligns with the operator’s gaze. The integrated camera eyeball and its operating system developed in this research hold promise for facilitating natural remote communication by enabling eye contact through avatars.

Reproducing the touching skills of Humanitude by robots is expected to reduce the burden of caregiving by humans and create more acceptable interactions for humans. In this study, we propose a robot hand reproducing the beginning of Humanitude’s touching motion. Specifically, we develop a robot hand with a mechanism following contact surface changes under the beginning of Humanitude’s touch method. To evaluate the robot hand, we confirmed that the robot hand satisfies the designed quantitative indices that represent the desirable touching motion by Humanitude professionals. Furthermore, we also conducted a subjective evaluation to measure the emotional arousal of a subject when touching the subject’s back using the proposed robot hand by contrast with that using a human-mimetic robot hand not considering the Humanitude’s touch method. We confirmed that humans could accept the proposed robot hand to feel more pleasant and less aroused.

Recently, research has been conducted on mixed reality (MR), which provides immersive visualization and interaction experiences , and on mapping human motions directly onto a robot in a mixed reality (MR) space to achieve a high level of immersion However, although the robot itself is mapped onto the MR space, the robot’s environment itself is not mapped, making it difficult to comfortably perform tasks that require detailed manipulation or manipulate objects that are difficult to see from the human perspective. Therefore, we propose a system in which the environment around the robot itself is also mapped on the MR space, and operations can be performed on the MR space, enabling immersive real-space operations.

Social robot platforms demand well-versed human behavior planning to be effective in their role of interacting with people. For a long time, despite the capability to perform a wide range of tasks, the lack of flexibility and effectiveness in decision-making on human-robot interaction is always the main issue of autonomous social robots’ severely limited real-world adoption. Social robot platforms nowadays are, though sufficiently sophisticated in behavior control with recent mechanical manipulation advancement, usually criticized for their uncanny and unnatural behavior planning. To improve this, we took a new data-driven approach to build a framework that learned from publicly available video data as an input with next-to-none annotation cost that could forecast and generate human-like facial behavior in an auto-regressive manner. The key idea was to break a long facial motion input into smaller sub-segments as a better cross-subject generalizable feature before feeding it into a Transformer-like cross-modal model. The goal was to create an open-source generative facial behavior component that could output to social robot platforms or other downstream tasks.

To investigate diary contents that elicit favorable impressions from readers, several diaries from the robot’s point of view were generated from the experience of walking with people, and impressions were compared for each content. The results revealed that mentions of the robot’s emotions and interactions during the walk increased the liking for the diaries.

As Well-being becomes increasingly important around the world, planetary health, which aims to coexist not only with people but also with nature, is attracting attention. In this paper, we will introduce new robotics developed with the aim of making natural objects the object of symbiosis instead of regarding them as the object of control. Specifically, we will describe the design of a leg-mobile robot and a switch for lighting, which were realized by using moss. We also conducted a questionnaire survey on the influence of robots made of natural materials on human sensitivity.

Communication of robots and AI increases in the future, Natural language communication of man-machine interface is a key technology. In addition, the Metaverse also requires automatic responses for avatars. In this research, we are developing a system that implements natural language processing in Python that realizes the precise local communication using avatars in Metaverse.

In this report, we propose a self-pose control for a mobile robot on an upward-facing LCD display. In the above environment, we have studied ”Foreign language conversation learning with direct method by using mobile robots” as previous works. But, we have a problem that an observer treats the same gestures of movements as the different gestures, if the mobile robots move unstably. In this report, we aim to limit accuracy of the mobile robot pose to 10[mm] and 8[deg.]. We achived the results of 5.2[mm] and 5.8[deg.].

In this study, we proposed a robot that requires the interaction putting it to sleep when co-sleeping with the aim of improving the quality of sleep, especially the time to fall asleep, for people affected by lifestyle changes. We summarized the intent and details of the functions to express physiological phenomena to be implemented in the robot and the functions to perform the interaction putting it to sleep, and also explained the scenario of using the robot. In the future, it will be necessary to verify whether or not the robot actually produces the effect of promoting sleep onset, and to explore the rules of the interaction process putting it to sleep.

We developed a five-degree-of-freedom (DOF) skill-assist arm with a passive joint to assist human upper limb motion while cooperating with its motion. In an assistance method, one of the arm’s three active joints is put into de-excitation, making it passive state. By rotation of the passive state joint, the arm detects the human motion, and an output point position is corrected on a target trajectory by controlling the angle of two active joints. However, because the skill-assist arm has a singular configuration, the arm must avoid it by switching the excitation of each joint. In the past, an actuator current was cut into zero in step-like to switch to the passive state. This current cutting caused vibration. Therefore, we proposed a method in which the current gradually reduces to zero. From the conducted experiments of the assistance using the proposed method, it was confirmed that the arm assists human motion with high precision.

Mobile robots are expected to be introduced into daily life in the future. Therefore, it is desirable for people around the robot to be able to coexist with the robot without feeling uncomfortable. Focusing on the situation where a pedestrian and a robot pass each other, if the pedestrian feels uncomfortable, the robot needs to change its behavior to improve its impression. In this paper, we propose a method for estimating the pedestrian’s impression of a passing pedestrian based on the measurement results of the pedestrian’s behavior just before the passing. The results of creating a discriminator that discriminates between comfortable and uncomfortable measurement results just before passing each other confirmed the existence of trajectories for which impression estimation is possible and trajectories for which it is impossible.

This study proposes a new approach for multiple users to view the same area simultaneously using a dual-camera system that combines an ultra-high-speed active camera and a wide-angle camera. A tablet with an inclinometer installed is used to control each user’s line of sight. This allows users to manipulate the tablet in order to view the magnified area they wish to see while controlling an ultra-high-speed active camera with multi-threaded gaze control at hundreds of frames per second. This system is considered to be an unprecedented system because it allows multiple users to share information while looking at their own desired viewing points when watching something remotely, such as a surveillance system, factory inspection, or sports game. In the realtime experiment, we confirmed the accuracy of tablet operation when there were two users by comparing it with automatic control, and then confirmed the movement with five users.

In many situations of remote control of mobile manipulators, cameras are used to grasp the external situation. In some cases, multiple tasks, such as moving and manipulating the manipulator, are performed in parallel in remote control of mobile manipulators, and in such cases, multiple cameras are mounted for each task. In this case, however, the operator is required to operate the manipulator while viewing multiple camera images as necessary, which increases the operator’s workload. Here, we considered the possibility of reducing the burden of reviewing multiple camera images by combining and presenting them as a single operation image. To realize this, we considered that there is a problem that the images necessary for the work may not be seen among the images of each other by merging the images. In this study, we examine a system to solve this problem.

Echocardiography is a diagnostic imaging method used by many medical institutions because the diagnostic equipment is small and inexpensive. However, damage to the examiner’s muscles and skeleton has become a problem in recent years, and research aimed at providing assistance has been conducted. We have previously developed Echography Assisted Robotics (EARs), a robotic system using 7-DOF manipulators that replicates a physician ’s ultrasound probe manipulation. Since manipulability of manipulators in cooperative diagnosis leads to physical fatigue of the examiner, it is necessary to place the patient in a location where the manipulator can move stably. Therefore, in this study, we investigated the effective range of motion of the EARs using Manipulability to determine the ease of movement of the tip of the manipulator. The results confirmed that the more the probe posture is perpendicular to the body surface, the better the manipulability. We proposed an ideal diagnostic environment for conventional diagnosis.

As medical technology improves, the number of upper limb amputees in Japan is increasing year by year. Although the demand for prosthetic hands is increasing, the percentage of electric prosthetic hands is low, accounting for only 2% of the total. We are conducting research to improve the accuracy of motion estimation and to simplify the training of using electric prosthetic hands. In this study, we used a piezoelectric sensor with a diameter of 0.5 mm, which is lightweight and easy to process, called an ultra-fine piezowire sensor. The accuracy of motion estimation using SVM was 92%, and the discrimination accuracy of the functional limb state was 100%. The accuracy of motion estimation using SVM was 92%. Significant differences were also observed in the muscle vibration when the arm movement speed was changed, suggesting the possibility of using muscle vibration to estimate the speed of change.

The objective of this research is to develop a novel human-machine interface (HMI) for individuals with paralysis in their upper and lower limbs. When using an interface to operate a control object, the paralyzed hand may cause delays or hinder the ability to reach the desired position. To address this issue, myoelectric signals, generated even during insufficient movement, can be utilized as input. By estimating the target position from myoelectricity, high-speed and high-precision control of the controlled object can be achieved through positioning control. The HMI can be applied to a wide range of applications, including wheelchair operation, operation of robot arms attached to wheelchairs, and work systems in the workplace. In this paper, we investigate a method for predicting the target position based on myoelectric patterns generated when the robot arm is moved to the target position. The findings of this study suggest the possibility of movement estimation through a certain relationship between the electromyograms generated during the attempt to move to the target position and the target position itself.

Knee flexions are relative motions between tibia and femur, which include rolling and sliding (rollback motion). Knee motions varies greatly among individuals. Conventional wearable knee assistive devices utilized hinge joints and generates nonnegligible mismatched motions under deep flexions. Therefore, we proposed a design method for a biomimetic knee joint (BKJ) that conforms to individual knee motion. A polycentric biomimetic knee joint (PC-BKJ) combining two gears with different radii is used in this study. In this paper, we developed a knee supporter based on the design method and evaluated the torque with the bio-artificial joint. As a result of an optimal design of the gear system, we adequately designed a pair of gears for a tailer made BKJ. Moreover, the torque of the optimized BKJ was mathematically evaluated. When the shape of the push link was changed without changing the spring constant, the maximum torque was 10.2 Nm, achieving the target of 10 Nm. Furthermore, the flexion angle at the maximum torque output was 132°, enabling assistance during deep flexion.

The number of patients with swallowing disorders is increasing according to the aging of society even though swallowing is an important role in the dietary process. In recent years, the simulator, named “Swallow Vision🄬”, made from medical images such as MRI and CT has been developed to visualize swallowing motion clearly. It enables us to understand kinesiology and analyze the motion of organs in swallowing. By using kinematic data obtained by this simulator and medical knowledge about swallowing, we have been developing a robotic simulator that has the potential to mimic human swallowing motion with same size and same time as human. The robot executes movements of tongue and pharynx to swallow food.

We aim to develop a method to actively sense an older adult’s health conditions using a dialogue with a robot in a way like a nurse or a caregiver. Our method estimates the health condition based on users ’responses which include response delay, volume, prosody, and other non-verbal information when the robot voice-calls to the user. The concepts of subsequent voice-calls by the robot, based on the estimated condition of the users, aimed to design encouragement maintaining or improving the user ’s health, including exercise, eating meals regularly, and sleeping, are generated with sensor integrations. This paper reports the voice-calling robot system constructed to test and verify the proposed method at home. The system is constructed to conduct dialogue as scheduled, change the contents of a robot’s voicecalling, and collect the sound data recorded during dialogue. We describe the outline of the constructed system, an operation confirmation experiment using the system. We also report a preliminary experiment and its result and future works.

In this study, we proposed an image processing method to detecting bed-leaving behavior. In our proposed method, feature vector is obtained by applying the wavelet transform to depth images that provide distance information between the camera and the target. By classifying the obtained feature vectors using Self-Organizing Map, our proposed method can classify three postures: lying down, turning over, and leaving the bed. We applied the proposed method to three targets and were able to classify three postures: lying down, turning over, and leaving the bed. The proposed method was also effective for unlearned data.

A model has been proposed to represent the relationship between ankle joint angle and ankle joint torque during passive motion using time-varying elastic modulus. The method assumes the viscosity term to be zero. This leads to modelling errors when the ankle joint is in passive motion at high angular velocities. In this study, a viscous term is added to this model. The validity of the proposed model was evaluated by showing that under experimental conditions with different angular velocities, the elastic torque obtained by subtracting the value of the viscous torque estimated using the identification model from the viscoelastic torque is almost identical.

In this study，the relationship between different assist patterns and gait changes were investigated through gait experiments under various assist conditions．We found that under an assist pattern，the overall muscle activity during the gait cycle did not increase，and the energy required for walking was reduced compared to the condition without assist．The reduction in muscle activity was observed in the hip flexor and extensor muscles，knee flexor muscles，and plantar flexor muscles．Furthermore，when using the motion-based assist，the activity of the plantar flexor muscles was significantly lower than that of C1，indicating its effectiveness in providing targeted assistance to the plantar flexor muscles．

Motor and sensory paralysis are the major sequelae of stroke. One of the causes of gait disturbance due to motor paralysis is the collapse of the internal model. We proposed an agnostic error as a rehabilitation method to prevent the collapse of the internal model and to improve voluntariness. We tested the method on 12 healthy subjects, and found that the addition of implicit error had a positive effect on motor change. The most effective results were obtained under the condition that agnostic errors were applied to all of the somatosensory FB (walking on the walking robot), visual FB (walking scene shown in VR), and load transfer FB (plantar vibration provided by the robot).

Since skin deformation can be obtained during passive movement by therapists, even stroke patients in acute who have difficulty with active joint movement, comparing skin deformations between passive and active movement and extracting the amount of active movement may be useful in assessing motor function. In this study, skin deformation was compared between passive and active movements based on the reconstruction error of a principal component analysis (PCA) model trained on passive movement data. Experimental results of measuring skin deformation of the upper arm during flexion and extension of the elbow joint showed that the PCA reconstruction error increased/decreased proportionally to the joint movement during active movement, and the average reconstruction error of active movement was larger than that of passive movement.

The purpose of this study was to verify whether training with the device improves reactive postural control, and to measure leg muscle activity during the standing posture to see if there are any factors that contribute to the improvement. Multiple measures of COP data showed that training improved reactive postural control in response to unexpected disturbances in the trained group. In addition, multiple measures of muscle activity showed that the dominant leg muscle activity changed when the unexpected disturbance came from the opposite side of the dominant leg, suggesting that training with WBTD seemly improvement reactive balance function immediately.

It’s reported that appropriate smells can improve psychological and physiological effects. Therefore, we developed Smell Wakka, a multi essential oil diffuser for rehabilitation support to achieve higher rehabilitation effects when used in combination with iWakka, a training and testing device for adjustability for grasping force. Smell Wakka can atomize six kinds of essential oils using an ultrasonic atomization method and a blowing fan, can be remotely controlled using Bluetooth. Moreover, we conducted performance evaluation experiments. The experimental results confirmed that two kinds of smells could be presented to the subjects by using the blowing fan.

Various control systems for robot arms using surface myoelectric signals have been developed. Abundant pattern-recognition techniques have been proposed to predict human motion intent based on these signals. However, it is laborious for users to train the voluntary control of myoelectric signals. In this study, we aim to develop a rehabilitation support system for hemiplegic upper limb. We develop training system to promote generation of voluntary surface myoelectric signals by manipulating virtual arm controlled via surface myoelectric signals. We propose a training scheme with gradually increasing difficulty level for virtual arm manipulation to evoke surface myoelectric signals. We make several training experiments on healthy subjects, and consider for possibility of clinic trial.

In previous studies, we developed a rollator with a rotatable chest pad. The rollator moves the user's upper body by swinging the chest pad so that they can easily step out. A novel treadmill-type gait training machine was designed and prototyped by applying a rotary chest support pad to a treadmill system. In this study, walk measurements were made using the developed novel gait training machine. Time series results and power spectrum of pressing force and lower limb motion were analyzed to investigate the effectiveness of the novel machine.

We are developing a hand rehabilitation robot with a remote power transmission．Rehabilitation robots need to be compact and lightweight so as not to hinder the movement of the user．In this study, we designed and manufactured a Remote Actuation System (RAS) to transmit a rotational motion from a distally located power source to the hand module work．In this paper, we presented improved system from our post system by adding a slider mechanism to get rid of a wire slack.

Wearable robotic systems for rehabilitation or motion assist require the users motion intention to control them intuitively. According to the construction of the human anatomy, some of the muscles contributing to different motions are located in inner layers of the muscle structure. Thus, when using muscle activity to estimate the human motion intention, currently available measuring techniques cannot adequately differentiate muscle activity based on different layers. In this study, by measuring the scattered light intensity from the inner layer muscles for a projected light source into the muscles, we measure the inner layer muscle activity. Initially, the source-detector distance was calculated using the Monte Carlo Simulation, followed by the prototype development. With the evaluation experiments by measuring muscle activity for finger flexion and wrist flexion, it showed different source-detector distances allows the measurement from different layers of muscles with different amount scattered light at each case.

Generating multiple patterns of grasping with different finger arrangements is an important characteristic of human hands during activities of daily living. Wearable robotic hands such as exoskeletons or prostheses, are used to supplement or substitute human hands in the events of rehabilitation, motion assist or amputation, respectively. To control them intuitively, it is important to understand the motion intention of the wearer. This has been a challenge with hand motions due to the availability of higher number of individual degrees of freedoms (DOF) of hand and the limited availability of information related to these motions, non-invasively. However, recently high-density surface electromyography (HDEMG) has proven to be effective in providing enough spatial resolution to access motion related information for different hand motions. Thus, in this study towards controlling a hand prosthesis we estimated the motion intention for 6 different finger motions by using time series data of HDEMG. Initially, HDEMG signals were recorded using a 64-channels electrode grid and using the root mean square values of the preprocessed HDEMG data, a 2D convolution neural network was trained to estimate the 6 selected different finger motions. Results demonstrated the proposed methodology can estimate the motion intention with an average accuracy of 85% and a highest accuracy of 92%.

We have developed a mechanical device that reduces the burdens on a caregiver’s lower back while they are in a bent-over position. We tested three kinds of prototypes during development. The first is a rotary model, which uses stoppers at the sides of the hipbones. The second is a truss model, which uses a triangular truss system in front of the user’s abdomen. The final prototype is a rope model, which utilizes the principle of leverage. The main unit of the final model on the back lengthens the moment arm of the user. This device can reduce the force used to keep a 25 deg bent-over position by 64.4%. This research reveals the most efficient area for the placement of the moment arm.

About half of the patients with spinal cord injuries in Japan have cervical spinal cord injury. Particularly, for patients with high-level spinal cord injury, it is difficult for them to search or pick up objects from the floor due to trunk dysfunctions. Recently, welfare robot arms of self-reliance supporting functions are being developed for them. In this study, we proposed an operating system that includes an eye in hand system with a touch panel interface for grasping the object from the floor. In the proposed method, the visual information of the target objects is shown in the touch panel interface. The patient could specify the target position of the robot arm by drawing a line on the target object on the interface.

This study aims to develop food printing technology using food powder. The composition of the food ink that is safe for nursing care food was determined based on a comparison with the lower limit of viscosity and the upper limit of hardness in Universal Design Food (UDF), as well as the discharge ability and shape retention after forming by a food 3D printer. In order to make the shape of the modeled object similar to that of an ordinary food, we used a scanner with a smartphone application to scan an actual dish, created the modeling data, and then created the modeled object.

This research aims to develop a food 3D printer capable of mixing and modeling multiple food ingredients in arbitrary proportions. One of the expected applications of food 3D printers is the provision of personalized foods. To provide personalized foods, it is necessary to print foods with ingredients whose nutrients and hardness are adjusted according to individual preferences and physical conditions, but no food 3D printer has been developed that is capable of blending multiple ingredients and printing foods based on data.

We created a food 3D printer equipped with multiple food ingredient tanks and a screw-type food ingredient extruder, where the ingredients meet at the nozzle portion, and by mixing two types of care food with different hardness in various proportions, we were able to mold food while adjusting the hardness of the food.

Individuals with neuromuscular or musculoskeletal impairments often face walking disorders, which significantly decrease their mobility and quality of life. This paper presents an approach to designing a mechanical walking assist that offers bilateral support for plantarflexion and dorsiflexion, supporting foot’s flexion and extension during walking by applying force at the ankle to adjust the walking motion and posture to a healthy pattern. To complete dorsiflexion, an offset slider crank mechanism lifts the foot, while a four-bar linkage assists plantarflexion by using the action-reaction force from the lever against the ground. The forefoot mechanical pressure sensor triggers both motions, and the heel sensor aids in releasing the lever for plantarflexion. Experimental results indicate that the apparatus provides assistance for both motions improving foot trajectories and ankle angle variation, leading to safer and healthier walking. This assist has the potential to improve mobility and quality of life for individuals with walking disorders, while also being applicable to other domains that require mechanical assistance, such as walking rehabilitation.

This paper presents an inflatable structure with a chain skeleton that can hold the human body. The inflatable structure is capable of tip extension and suppresses sliding against the human body. The chain skeleton enables switching stiffness and provides a secure hold. We developed a prototype that integrates these elements into a compact system.

A water flow pump like a jellyfish using two finger robot gripper has been designed. A strong two-dimensional flow, that intermittently pumps water mass downward, has been generated by opening bell and closing bell. In this study, the characteristics of the flow depending on the position of the flow guide plates inspired by a mouth-arm of jellyfish were investigated. Dye visualized flow patterns around the guide plates showed that the induced flow to downward from around the bell was affected by the position and installation angle of the guide plates. Especially, the angle of the guide plates was important in preventing the generation of backflow.

We have been developing motor-tendon-joint series connection type biomimetic actuators to make robots act jumping, running and dancing. To that purpose, we are now developing an actuator with O-ring tensile tendon structure. In the previous paper, we succeeded to apply the energy of landing shock for the next jumping energy in the shallow knee angle area. We also found that the condition of the landing knee angle around 132 degrees is the best angle to use landing shock to the next jumping. Therefore, in this paper, we investigated how to set the landing knee angle to 132 degrees during continuous jumping. By using this control scheme, the amount of jumping increased 2.1 times from 180mm at the first time without using landing shock to 380mm. We also examined a landing sequence to achieve soft landing with less bounce. The best landing knee angle for the soft landing is 95 degrees.

In this paper，we propose a method for moving under a load, such as a human body, using a flexible "Wavy-sheet" actuator that generates continuous traveling waves. Wavy-sheet is pneumatically driven, and its flexible and thin-walled structure makes it suitable for sneaking into confined environments. However, the tube structure attached to the side of the sheet causes strong frictional forces, making it difficult to move under human loads. In this paper，We clarify the moving principle of Wavy-sheet and its velocity. Also, we propose a new type of Wavy-sheet with a built-in tube, and verify the effectiveness of the proposed method by experiments in a flexible gap under human body load, which has been difficult to move in the past.

This study seeks to realise human-like motions by mimicking the number of ligaments, the kinds of ones, and shape of femur and tibia, and we proposed that there is a possibility that they improve the degree of human-like motions by mimicking them. We conducted experiments to ensure that ligaments constrain the motions as humans do. We confirmed that slipping and rolling motions, prevention of overextensions, and yaw rotation were realised. Throughout the experiments, we indicated the possibilities of collateral ligaments’ contribution to prevent overextension. We formed two hypotheses to validate what extent these motions affect on humanity, and conducted a survey. Consequently, the present of both motions did not have significant differencies in it.