Commercially available multi-degree-of-freedom prosthetic hand designs are not functional compared to conventional 1-DOF hands. We hypothesized the thumb mechanism to be deficient and designed a single motor driven 7-Link closed-loop mechanism to oppose the thumb to the index finger in multiple prehensile forms along a single motion locus. The mechanism was modeled using the DH method and set the target motion. The target motion was set to move the thumb’s distal phalanx opposing, at first to the side of the index finger, and then to the and distal pulp of the closing fingers. The effect of the levels of link length a and link joint axis angle α’s combinations were simulated and elevated with the objective functions: angle between position vectors of initial and final distal phalanx coordinates. The combination a₃=30, a₄=95 mm, and α=100 deg. was found to be the most optimal design parameters for thumb 7-Link closed-loop mechanism.

A surgical instrument developed as a substitute for the conventional retractor for use during surgery is described. As for the conventional retractor, it was necessary to manipulate all the knobs for joint fixation of each joint to keep its position and posture. The new type of proposed retractor has a mechanism that can lock and unlock all joints simultaneously by turning just one lever. The mechanism installed in each joint consists of miter gears and universal joints. The retractor which has a multiple D.o.F has a high controllability as a manipulator. The position and orientation of the retractor can be changed by movement of each link. Each link is connected by non-driven joints at its both ends, and all links are driven by manual. Since the actuators are not installed at any place in the retractor, it is washable. With this mechanism, it is possible to hold the situation of shape of the target part during surgery. Thus, the new instrument can increase performance level and efficiency of surgery.

A surgical instrument developed as a substitute for the conventional forceps for use during surgery is described. The new type of proposed forceps has a serial multiple link mechanism that consists of an oblique circular cylinder. The mechanisms installed in each joint consists of elongated springs, lock mechanisms with small springs, a lock-release mechanism, a release mechanism, a motion control mechanism, etc. Especially, we explain the analysis of characteristics of the small spring which is installed on lock mechanisms and the elongated springs. The position and orientation of tip on the forceps can be controlled by changing the rotation movement of each adjacent link. Each link is being connected by oblique circular cylinder shaped non-driven links and joints, and all links can be driven by the conduction force generated through the spring. Actuators are not installed at any places in the forceps manipulator. Also, the manipulator must be washable for use in surgery. It is possible to manipulate the forceps, cut out, and suture the target part in the narrow space while changing the shape of its whole part avoiding essential parts of the tissue as minimally invasive surgery in order to increase performance level and efficiency of surgery.

Adjusting myoelectric sensor sensitivity of a myoelectric hand controller requires experience and knowledge. Malfunctions due to inadequate sensitivity adjustment can reduce the reliability of the prosthetic hand and induce interruption of its use in daily life. Therefore, the objective of this research is to develop an automatic sensitivity adjustment function. RT method is applied to accomplish resistance to disturbances with efficient open/close discrimination to the muscle activity level. Therefore, in this paper, we designed a myoelectric prosthetic hand controller with Robust Parameter Design. By selecting 7 control factors’ level combination, confirmation test results showed that the open-operation system had a signal-to-noise ratio of +25.9 % and a sensitivity of -16.2 %, indicating reproducibility of the optimum conditions. Similarly, the closed operating system had a signal-to-noise ratio of +88.2 % and a sensitivity of -32.6 %, which did not support the reproductivity of the optimum condition. However, the optimal condition of the closed operating system had better signal-to-noise ratio 12.8 db over the comparison condition.

The multi-joint linkage body powered finger prosthesis has a linkage mechanism that converts the flexion and extension movements of the residual finger end into flexion and extension movements of the prosthetic finger. However, as the motion magnification ratio increases, the prosthetic finger does not provide sufficient grasping force at the fingertip. Therefore, the research objective was set to improve the grasping ability by implementing the prosthetic finger with a hyperextension mechanism. A hyperextension mechanism consisting of a 3-joints link mechanism and an elastic band was designed. To evaluate the stability of grasping, fingertip force was measured when applying rotational torque to the model finger with a conventional prosthetic finger setup and a finger with the hyperextension mechanism. As a result, the prosthetic finger with the new mechanism showed 6 and 13 % greater fingertip force in the shallow-flexed and flexed positions, accordingly. However, the force decreased 23 % in the deep-flexed position. Modification of the elasticity design of the band should be made to adjustment.

Understanding of the musculoskeletal structure of the lower limb to realize human gait will provide important knowledge in the field of rehabilitation and other fields. In this research, we aim to construct a musculoskeletal robotic gait simulator to reproduce human gait by driving some muscles. This research is not a study of a biped robot, but a biomechanical study of human gait. In this paper, we refered to human gait as rocker function and attempted to construct a gait by focusing on the forefoot rocker among the four rocker functions. Forefoot rocker was reproduced by driving the gastrocnemius muscle at the same timing as humans. However, knee flexion occurred because the gastrocnemius was driven strongly. The rectus femoris muscle was adjusted to accommodate this. This resulted in the development of a forefoot rocker and a decrease in the amount of center of gravity drop. In addition, the robot was able to generate a subsequent gait, which is reported here.

Maneuver of flexible endoscopic operation requires a difficult precision operation for a long time period. So, it is expected that a robotic system for flexible endoscope handling. We developed the Endoscope Operation Robot and a micro robot arm for endoscopic dissection. The robot equipped bilateral haptic feedback, which can represent reactional force to operators in forward and backward movement and rotation of an endoscope. The micro robot arm for dissection operation has 5 degrees of freedom which independent of the endoscope operation, and endoscopists can easily manipulate the robot arm with a pen device. Experimental endoscopic submucosal dissection (ESD), one of the treatments for early-stage gastrointestinal cancer, was performed with our system. All 7 cases of ESD were completed with en bloc resection, and mean dissection time of our system were 1.88 min/cm² or 0.67 cm²/min. In comparison of mean dissection speed of our system with other systems, our system was the quickest. We showed that our system can perform ESD with high efficiency in this paper.

In this paper, we propose a technology for diagnosing the state of objects for automated inspection of facility infrastructure. As part of efforts to link and integrate non-destructive inspection data, we automatically estimate the deterioration state of steel drums that store waste based on image recognition. Specifically, the area and type of rust on the steel drums surface are automatically estimated by using semantic segmentation.

Most public infrastructure equipment in use today was built during a period of high economic growth and is rapidly aging. However, inspection workers have been decreasing and tools for inspecting buried infrastructure (e.g., handheld ground-penetrating radar) require a great amount of labor and time. Thus, autonomous mobile robots are required to improve efficiency and labor costs. High self-position estimation accuracy is required for autonomous movement and for specifying a position by comparing the coordinates of objects buried underground with information from a ground-penetrating radar. In this report, we describe a method using road surface motion images as a self-position estimation applicable to underground burial exploration robots. To compensate for accumulated errors, we propose an algorithm that records feature points at regular intervals. Experiments show that the error is corrected by using the accumulated feature points.

We developed the cyber-physical system(CPS) which can achieve stable and high quality laser welding process by automated control of welding conditions with AI judgement of welding states and simulation-based process database. The system which has monitoring, recognition, judgment, and process control achieved autonomous laser welding. In this presentation, we will introduce some practical examples for high quality welding by CPS and its application for process quality assurance.

In the periodic inspection of Photovoltaics, since the inspection target area is vast, it is desired that labor-saving and low-cost maintenance management using unmanned aerial vehicle (UAV), artificial intelligence (AI) and so on. Many inspection systems for the upper surface of solar panels have been developed. However, there was no inspection system that could detect various defects under the solar panels. Therefore, we have developed an automatic inspection system for the underside of the solar panels. By suspending the camera from the UAV to take pictures, stable shooting under the solar panels was realized in response to various ground surface environments. In addition, by using an AI model created from normal imaging data, we succeeded in detecting various defects.

The demand has now arisen for nondestructive methods allowing spot welding points to be automatically and swiftly inspected with high accuracy. We have developed a spot welding inspection robot integrating Matrixeye^TM and robot control techniques, which makes it possible to adjust the tilt angle of the ultrasonic probe. In this paper, we introduce this robot system and method to inspect spot welding without couplant, which enables to reduce the inspection time and to eliminate set-up operation to wipe out couplant. Also, we suggest a cyber-physical system by the spot welding automatic inspection system, which can contribute to improve the quality of the spot welding process.

This paper reports on an acoustic resonant imaging technique that visualizes the acoustic properties of a thin film formed on the substrate. A photoresist film was formed on the top surface of the Si substrate, and the echoes reflected from the bottom surface of the substrate were recorded to create the acoustic image. By analyzing the frequency domain data of the echoes, the distribution of acoustic impedance was successfully visualized.

Aging infrastructure has become an issue of social importance and there is a desire to use information and communication technology such as Internet of things (IoT) sensors to efficiently maintain and manage such infrastructure. This paper proposes a novel, ultra-low-power acoustic emission (AE) amplitude–frequency distribution sensor for monitoring the structural health of infrastructure. AE is a weak elastic wave generated by cracks, rubbing, and other damage in structural materials. The relationship between AE amplitude and frequency distribution is known as Gutenberg–Richter’s law, and its slope (the b-value) is considered an indicator of material integrity. The proposed sensor was designed to monitor changes in structural integrity by detecting b-value trends over a long period. We focused on two key points for application as an IoT sensor: ultra-low power consumption and long-distance wireless transmission. The wake-on AE algorithm, which wakes the sensor system when AE waves are generated by damage, enables AE sensing while consuming very little power. In addition, the unique analog–digital cooperative circuit technology successfully extracts AE features with less sampling. Meanwhile, statistical analysis at the sensor edge contributes to reduced data rates and allows for the use of low-power wide area networks for data transmission.

Road bridges are required to be inspected once every five years. Therefore, it is expected to improve the efficiency and accuracy of inspection technology of road bridges. In this study, we develop a vibration sensing actuation device that solves these problems. The device constructed from an IoT sensor module, an electrical relay circuit, a white noise generator and an electrodynamic exciter. The applicability of the device was evaluated using a experimental beam to see if cracks could be diagnosed. As a result, it was implied that the probability density of acceleration changed depending on the presence or absence of cracks.

Road bridges are required to be inspected once every five years. Therefore, it is expected to improve the efficiency and accuracy of inspection technology of road bridges. In this study, we develop a vibration sensing actuation device that solves these problems. The device constructed from an IoT sensor module, an electrical relay circuit, a sinusoidal signal generator, amplifier circuit and an electrodynamic exciter. The applicability of the device was evaluated using experimental beams to see if cracks could be diagnosed. As a result, it was implied that the probability density of acceleration changed depending on the presence or absence of cracks.

We have proposed a soundness evaluation method for RC slabs. In this method, when AE generated on a road surface while a vehicle is running is observed using an AE sensor installed on the underside of the RC slab, the more the RC slab deteriorates, the more the propagation of elastic waves is hindered, and the density of the observed AE source becomes thinner. It was confirmed that the temperature is a variable factor of the index when AE source density is used as a soundness evaluation index. Verification of the fluctuation factor by elastic wave propagation analysis revealed that the temperature dependence of the elastic modulus of the pavement affects the detection waveform of the AE sensor on the bottom surface of the RC slab. In addition, consistency with the fluctuation of the soundness evaluation index in the AE measurement results was confirmed.

In this paper, a vibration test and analysis of cracked beam is performed focus on the change of mode shape. In previous investigation by Research Association for Infrastructure Monitoring System, many field tests are conducted using deteriorated bridges which has multiple damage pattern. As the result, the experimental data of mode shape change which is associated crack occur are obtained. However, the mechanism why does it change the mode shape is not understanding. In this paper, the vibration test and analysis are conducted to obtain the fundamental data of cracked beam. Especially, the experimental modal analysis and theoretical analysis based on cracked Timoshenko beam were conducted in case of the metal material.

Pipe wall thinning management operations have been required to be strengthened in the wake of previous pipe rupture accidents. As the number of target locations increases, it is necessary to manage a vast amount of data on pipe wall thinning. Therefore, Pipe Wall Thinning Management System will improve the accuracy of pipe wall thinning management by centrally managing a vast amount of related data, analyzing the trend of thickness reduction, and formulating appropriate inspection plans.

In this study, we examine the time-reversal imaging for guided wave inspection in a water pipe. In general, the guided waves exhibit multimodal and dispersion characteristics. Time-reversal imaging is one of the promising methods because the dispersion characteristics can be taken into account automatically. Green’s function, required in the process of time-reversal imaging, is calculated using normal mode expansion method for guided waves. The normal-mode solutions are obtained from the result of the semi-analytical finite element method.

Excitation of local defect resonances (LDR) using a relatively high-frequency vibration to estimate various quantities of defects has been reported. To investigate the characteristics of the LDR, numerical modelings of LDR were conducted in this study. The finite element method (FEM) and boundary element method (BEM) for a model containing a crack were implemented to obtain the LDR frequencies. Since the FEM outputs not only LDR but also whole domain resonance, we found it troublesome to extract the LDR solution. On the other hand, the BEM based on the crack discretization provides only LDR solutions caused by the crack.

In order to develop a overhead crane that can provide high safety and productivity, the authors developed a technology for rope tension detection and load estimation of payload. The technology can be applied to detection of overload and lifting off from ground. Multi-task learning using neural network was adopted to estimate the output of induction motor with nonlinear properties. The driving signals were input to the neural network, and the rope tension state and the load were estimated at the same time. The authors proposed a method to create teacher data by using slip value of induction motor to determine the rope tension point based on actual load. As a result, it was clarified the proposed method is effective for evaluating the rope tension point based on the increase of actual load. In addition, the accuracy of the learning result that was learned by the proposed method was improved from the previous method. As for the accuracy of the proposed method, the deviation of the rope tension detection error was less than ±0.5 second, and the load estimation error was 2% of the maximum load.

Picking robots are required to pick and place objects, which have a variety of physical properties such as size, weight and flexibility. This paper presents a robot hand control method for picking robot which can adapt physical properties of objects. In our method, robot hand control system calculates safety ratio for grasping an object by integrating object information and contact information. Object information includes weight from object database, shape from an RGB camera, position and pose from planner for grasping. Contact information includes measured suction pressure and suction area which is calculated based on a measured value of pressure sensor. This safety ratio is used to determine whether suction status is sufficient to lift robot arm up and also determine acceleration of robot arm. To evaluate our method, pick and place tasks were executed for various objects which are difficult to carry stably with multiple suction pads. As a result, the success rate for picking and placing objects was increased, compared to controlling hand without our newly introduced safety ratio methods.

In this paper, an object category method and grasping method is proposed for housework assisting robot. Multiple kinds of objects that are trained previously by CNN (Conventional Neural Network) can be recognized and categorized by the recognized class names. Other unknown objects are categorized by the similarity of object features. The recognized objects of same category are grasped by the robot and moved to the specified places. The grasping postures are decided by the detected spatial shapes of objects.

In tactile virtual reality (VR), there is a limit to realizing the real objects’ textures only by hardware. We focus on Velvet Hand Illusion (VHI), in which a person rubs his/her hands together on either side of two wires stretched in a parallel to produce a sensation of rubbing a smooth virtual surface like velvet on the palm. We conducted a psychophysical experiment by changing the wire diameter to investigate the relationship between the smoothness of virtual surface and wire diameter, and one between the thickness of the virtual surface and wire diameter. Based on the experimental results, we newly proposed the closure factor defined by the two hands as palms together factor. Since a virtual surface doesn’t occur with only one hand in VHI, this factor is essential for the induction of VHI in tactile Gestalt, which cannot be observed in sight or hearing sensations.

In this study, a general-purpose saccade measurement method was investigated using multiple webcams to obtain high-speed pseudo-facial images. An eye-tracking system was constructed using a high-speed shooting method, in which the exposure timing changed when the cameras were reconnected, and the camera reconnection was adjusted such that the difference in the exposure timing was equal to the target shooting speed. In the experiment performed to verify the high-speed shooting method, although an increase in the variation over time was observed, the shooting speed was quasi-high when multiple cameras were used. In the saccade measurement experiment, the error rate of the saccade peak velocity was observed to be approximately 28 % when the stimulus movement was 10° or more, which was significantly lower than the error rate when a single camera was used. These results suggest that saccadic motion with a large shift of viewpoint, such as when viewing advertisements, can be captured using an inexpensive device for general use.

One of the factors of good explanation is that the listener's understanding is confirmed. Therefore, we focused on ocular information as an objective index to grasp the comprehension status. It has been suggested that blinking affects psychological load. Pupil diameter has been suggested to affect the cognitive load to comprehend sentences. In this study, we evaluated the influence of ocular information on the degree of comprehension during listening to sentences composed of unknown and known words. In addition, a subjective evaluation questionnaire was administered to listeners to evaluate their comprehension of the sentences they heard. The results showed that the blink rate after listening to a sentence was significantly higher and the pupil diameter was significantly mydriatic when the comprehension level of the sentence was low. Furthermore, the accuracy of estimating the listener's comprehension level from their eye information was estimated by SVM which is categorized in machine learning. According to the result, we found out that the accuracy of estimating the comprehension level was 70% and that of estimating the ease of comprehension was over 60%. Therefore, it is suggested that it is possible to estimate the listener's comprehension level by using their eye information.

In recent years, there has been a growing interest in using silent speech to generate speech and recognize speech content for various applications in the fields of medicine, human interaction, and entertainment. Gaddy et al. have developed a machine learning model that can generate speech from muscle activity (using electromyography (EMG)) with 64 % word recognition accuracy for one English speaker. This study examines the possibility of applying this approach to Japanese speech synthesis by fine-tuning a machine learning model with English data, investigating the useful EMG locations that have not been measured in previous studies, examining the feasibility of learning the EMG data from non-speaking individuals paired with speech from other people, and evaluating the effect of using phoneme-balanced sentences for improving the word recognition accuracy. The results of this study suggest that speech synthesis in Japanese is possible with a limited vocabulary, and that fine-tuning with English data improves the accuracy by 40 % relative to not performing fine-tuning. Adding more EMG locations, particularly in the neck (styloglossus and hyoglossus muscle groups), improves the word recognition accuracy by 60 % relative to not adding them. It has also been observed that it is possible to generate speech by learning the EMG data paired with speech from other people, and the usage of phonemebalanced sentences for data creation has been found to be useful. It is expected that the word recognition accuracy will improve with an increase in data in Japanese. Furthermore, it is expected that this technology will be able to generate speech in Japanese with unrestricted vocabulary.

With the development of smartphones and ease of access to high-speed internet, many people now spend a lot of time on their smartphones and tablets, especially for watching videos. In recent years, many people tend to watch dramas and movies at increased speed to save time. Flickering and near point distance were measured as objective indices in the experiment, and LF/HF was used as an auxiliary method to analyze examinee’s fatigue. The "subjective fatigue test" was used as a subjective index to evaluate human subjective fatigue. Therefore, the purpose of this study is to investigate the effects of 1x and 2x playback speed on human eye fatigue during video viewing. The result shows that although there was no statistically significant difference in the experimental results, it can be inferred that there was a downward trend in flicker change, which expresses mental fatigue at 2x speed, and that there was no change in near point distance, which expresses regulatory fatigue, at either 1x speed or 2x speed.

To realize the technology for recording and reproducing procedures performed by highly skilled technicians, we develop a nail color sensor that can be worn on a human fingertip to record tactile information. The color of the nail changes depending on the force applied to the finger cushion and the posture of the finger because the positional relationship between bones, capillaries, and nails changes due to the elastic deformation of the finger cushion. In this study, we develop a nail color tactile sensor that can estimate force information and finger posture information from this change. As a result of the sensor performance evaluation, it is shown that the posture and fingertip force can be estimated simultaneously.

It has been shown that two kinds of phenomenon called kinesthetic illusion and tonic vibration reflex (TVR) are induced by providing vibration stimulation to tendon of muscle. Here comes to a possibility that using kinesthetic illusion to generate sensation of movement and/or using TVR to achieve the movement. This research focuses on the generation of kinesthetic illusion of shoulder external rotation by stimulating infraspinatus. In order to increase the kinesthetic illusion effect, the study focuses on frequency of the vibration stimulation and simultaneous skin stretch. It has been shown that 80Hz generate the most effective kinesthetic illusion. Furthermore, the result shows that simultaneous skin stretch increases the amount of kinesthetic illusion.

For body movement instruction, various wearable devices have been proposed. However, their users sometimes suffer from a sense of restraint because of the structures of such devices. Therefore, the authors proposed a novel instruction device with a pendulum that is attached to a tool, not to the user’s body, that is, this device does not restrain the user’s body movement. As the center of gravity of the developed device that is attached to the tool translates by swinging the pendulum, the user is expected to perceive a guiding force as if the user’s body is moved by an external guiding force. In this paper, in order to shorten the reaction time until the user perceives the presented guiding force generated by the pendulum, the effective configurations of the pendulum device, i.e., the angular velocity of the pendulum rotation, the mass, the length and the rotation angle of the pendulum are experimentally investigated.

Patients who temporarily use crutches may have secondary accidents such as falls due to lack of crutch walking skills. To reduce such a possibility, a device that presents a sensation to be falling was proposed. By training the patient with the device, it is expected that the possibility of a secondary accident would be reduced. A step of a crutch walk motion can be divided into six phases, i.e. , early, middle and late of swing phases and those of stance phases. If the phase in which a crutch user tends to lose the bodily balance easily is identified, a training can be provided specifically for that period. In this paper, we investigated the crutch walk motions through experiments, and its result shows that a training had better be provided for the swing phase.

Due to the revision of international organization for standardization about collaborative robots, without the safety fence work people in the workspace with the robots. Even in human society, we encounter more opportunities with mechanical systems such as advanced driver assistant system. In order to interact with the system properly, the field of human machine interaction has been studied. It is concerned that the assistant system provides people with so much convenience that they use the system too much. The result was assumed to lead to an accident because the system works by no means. The reason of the overuse was thought to be reliance as the internal factor. Previously we estimated the reliance on assist system. In this paper we examined the relationship between the performance and reliance and considered its validity for performance improvement. The result showed that work performance was changed according to the assistant status as well as human reliance.

In recent years, lack of sleep has been a serious problem in Japan. Lack of sleep may cause work performance loss. One of the solutions to lack of sleep is napping. Napping has the effect of inhibiting decrease in work performance due to lack of sleep. However, the wrong way of napping may cause sleep inertia on awakening. Sleepinertiaisstrongdrowsinessorheadacheafterawakeningand, leadstoaworkperformanceloss. Inthisstudy, we focused on the relationship between sleep and temperature and, we aimed assisting to awakening with the thermal presentation to reduce sleep inertia. The experimental results suggested that heating thermal stimulation was effective for subjective sleep inertia, while cooling thermal stimulation was effective for objective sleep inertia. However, autonomic nervous activity and Lorenz plot suggested that there were individual differences in sleep state transitions between heating and cooling thermal stimulation.

Technological advances have made automation and assisted systems more prevalent in society. However, it is difficult to emulate the flexible and delicate movements of humans and the intuition of skilled operators. For this reason, some control methods include the estimation of operating intentions and skill evaluation in human-machine cooperation (HMS). However, it is necessary to consider the reaction of humans to operation assistant in the situation. In this study, we focused on the influence of the machine approach on humans in HMS and proposed a variable admittance control based on a system architecture for ”Sensing The Atmosphere”. We evaluated the positional operation performance with relation to the natural frequency and used the steepest descent method to search for the optimised natural frequency. First, we tested the proposed method for operation assistance. Second, the relationships between machine impedance and human operating performance were measured. Finally, we considered methods for modelling changes in human operating characteristics in relation to the machine impedance.

A household disaster map is required as a countermeasure against earthquakes. In particular, evacuation is difficult in a messy crowded indoor space. Therefore, the visualization of areas that are likely to hamper evacuation is important. This study focused on cleaning robots. Cleaning robots generate environmental recognition maps to control their movement. This study proposes a system that detects obstacles impeding evacuation in households. The system used an environmental recognition map generated by a cleaning robot. The map generation algorithm was based on image processing and stochastic virtual pass analysis based on a pseudo cleaning robot model. Stochastic virtual pass analysis was performed to track the coordinates (the virtual pass of the robot model) inside the room. Furthermore, the proposed system was tested in a laboratory. Moreover, the simulation of layout design changing was conducted.

When a robot system teaches human, the adaptive method is required for teaching suitable to each person. Such as “Assist as needed” or “Gradually decreasing assist.” On the other hand, humans attempt to teach others to adapt to each person through interaction. Due to interaction, a trainer’s input is different from regular task inputs. The analysis result of this different input showed that Trainer teaches Trainee by switching teaching term and checking term of learning progress. To analyze the difference checking intervals, we propose a task that can switch teaching and checking. The results showed that conditions checking once in four times has a bad effect on initial learning. And, conditions checking once in two has a bad effect on the latter term of learning. Thus, the results suggested that conditions checking once in three can teach stalely stable for teaching.

Standing balance is one of the most important elements for the health. Falling accidents, which have increased in recent aging society, are deeply related to standing balance. Therefore, in recent year, many people have been doing research on balance. Currently, the most major research is investigating the relationship between tactile stimuli and balance, but they have a disadvantage to require some large and complicated devices. Thus, in this research, we will attempt an approach using auditory stimuli for research that measures changes in balance. By the way, the sound have various elements. Therefore, in this experiment, we focused on the relationship between sound localization and center of pressure and the results are evaluated by sway path length and sway area as in other balance studies.

Social Robots are growing in plurality. Therefore, We need to take into consideration that the human operating the robot may be a caregiver or other unskilled operator when we design our service robots. In recent years, (Conversational Agents: CA) is received attention from many domains. The problem is that these systems cannot estimate the details of the speaker’s expression. In a previous study, we proposed estimating intent in more detail. However, many issues remain in the speaker prediction model that needs to address. Therefore in this paper, We perform the experiment to investigate the relationship between speaker expression and listener prediction intention. Results suggested that listeners rated each speaker individually.

This study develops the robotic IV pole for reducing the risk of falling. The risk of falling caused by tripping over the conventional IV pole while walking can be reduced by having the IV pole autonomously move side by side with the user (the user does not touch the IV pole). The Robotic IV pole comprises omni-wheels, Light Detection and Ranging (LiDAR), Raspberry Pi, and Arduino Due. The proposed method is a controlling method for following the user’s travel direction and keeping distance between the robotic IV pole and the user, using a point cloud of the user’s both feet that can be obtained by LiDAR. An experiment confirms whether the proposed algorithm can follow the user’s travel direction and keep the distance between the robotic IV pole and the user. As a result, both algorithms work well.

In physical Human – Human Interaction (pHHI) research, actuators are used to construct physical interactions. Various types of actuator control methods have been adopted by researchers, such as Position – Position (PP) control, Admittance control, and Bilateral control. However, how different control methods affect human operation and interaction characteristics has not been clarified. This study aimed to clarify whether pHHI depends on the control method. In the experiment, a Leader – Follower system was constructed using four control methods: PP control, Admittance control (by PD controller), Admittance control (by Acceleration Based Controller), and Bilateral control, and subjects executed the cooperative task. The objective evaluation showed that the admittance control resulted in a more significant angular error between interfaces than the PP and Bilateral control. Also, the subjective evaluation results showed that the admittance control decreased both the naturalness and quality of the interface perceived by the user. Admittance control has been unstable depending on virtual dynamics, which may have induced the drop observed in the subjective evaluation scores.

The purpose of this study is to improve the English pronunciation of Japanese people by focusing on the English pronunciation method that Japanese people have difficulty with, and by analyzing various approaches such as EMG analysis, image analysis, and voice analysis in speaking English. Experiments were conducted by measuring the EMG signals around the oral cavity during the pronunciation of English vowels, video of mouth movements, and measurements of the pronounced voices. The experimental results were analyzed by each method, and the possibility of vowel identification was verified. The results showed that it was possible to identify some vowels, but not all.

This is an activity report of the industry-academia-finance collaboration for manufacturing DX that has been jointly planned and promoted by Musashino Bank and Toyo University and developed in Saitama area. This academy is based on lecture series for collecting new technology information, exercise series of basic technology mainly concerning on IoT and the introduction of new digital technologies into the participants’ own company. The academy provides not only knowledge but also the collaboration place where engineers in different companies can share various information each other. The academy lasts for four years until now, and continue to the new series.

In recent years, the participation of robots in human society has become more practical. In order for robots to adapt to the society, we think that robots should not interfere with human activities. Human attention is one of the factors to estimate human intention so that robots can work unnoticed from people. Since humans visually tend to obtain the surrounding information, human attention distribution can be determined by analyzing eyes’movements. Then, Avoiding the estimated human attention distribution could contribute to the unnoticed behavior. In this paper, we proposed a method for estimating human attention distribution in three dimensions. The parameters of the proposed method include the idea that the spread of human attention distribution varies depending on the speed of the gaze shifting．An experiment was conducted to verify the calculated values of 3D Attention in places that the subject unnoticed. The results suggested that the proposed method can identify locations that are difficult to notice.

Today, with the development of robotics and network technologies, robots that can be remotely controlled based on images from cameras have been developed. Since these robots are difficult to operate due to their remote-controlled nature, researches being conducted to reduce the difficulty of operation by implementing operation assist System. However, depending on the timing of the operation assist, it may interfere with the work or induce hazards. Therefore, in this research, based on the assumption that such unwanted assists will occur, aims to improve the reliability of operation assists by eliminating unwanted assists through assist selection. When operate a remote-controlled robot, the operator has to move its line of sight in order to check the camera images. Therefore, we have studied a method to estimate the distribution of attention during operating from eye gaze and to select assists according to the degree of attention. First, we prepared a robot equipped with a control support system in a virtual space, and asked subjects to remotely control the robot. At that time, we selected assistants using the proposed method and evaluated their task completion time and the number of mistakes they made. As a result, it was confirmed that unnecessary assists were eliminated, but there was a tendency that some necessary assists were also eliminated.

This paper investigates the optimization combination of search methods for multi-drone considering influences of the wind above sea in the sea rescue through simulation experimentations. This study employees adaptive optimal guidance law (AOGL) as control law for drones in order to prevent that a drone drifts by the wind above sea. As simulation results, when we consider the wind above sea, the discovery ratio of victims by multi-drone decreases. However, the average discovery time of victims was same level regardless of whether considering the wind above sea.

Man-hour tallying at manufacturing is often performed manually by workers, but there are concerns about input errors and the time and effort required for inputting man-hours. To improve this problem, several studies about automation of man-hour tallying have been conducted like include indoor positioning using tags and antennas, and methods in which workers wear various sensors. However, these methods have disadvantages in that they place a heavy burden on workers or the systems are large and expensive. In this presentation, we will present a technology and development of a simple man-hour counting method using a camera installed on the ceiling without placing a burden on the workers.

The development of intelligent tires that employ sensors to monitor road surface information and tire conditions has been underway for realizing safe and secure mobility for society. This paper describes a wear estimation method for tires using piezoelectric films arranged in parallel. We compared the outputs of multiple piezoelectric films between the new and worn tires. The ratio of the peak interval time to one tire rotation period(time ratio) and the ratio of positive voltage to negative voltage (voltage ratio) were analyzed. The position where the voltage ratio is large shifts in the direction of the tire side due to wear. There is no difference in the time ratio in the central part of the tire before and after wear, but it decreases in the lateral part of the tire. These are due to the fact that the contact patch width widens due to wear, and the sidewall section begins to support the load.