This paper describes structural design and circuit design of “Net-Structure Proximity Sensor” for mobile robots, especially snake robots. Requirements for the sensor is to refrain for inferring any motion of the robots and to be easily added to the control systems. The size and shape are determined to prevent both contact to the road-surface and interference to the movement range of the servomotors of the robots. The sensing and amplifier circuits are designed to enhance the sensitivity for attitude errors against the road-surface and to be driven by single power supply, which makes it easy to drive the sensor by the Li-polymer batteries and to handle the output signals by the microcomputers.

A novel method to compute contact forces based on contact volume with polygonal approximation of Coulomb's friction law in forward dynamics simulation is proposed. The contact forces applied in the volume is modeld as the resolutant 6-axis force of the distributed linear forces on the contact plane. The constraint condition on the 6-axis force derived from the unilateral and frictional constraint of each distributed forces is approximated to the linear programming problem with polygonal approximation.

Multirotor helicopters are expected to be utilized for various tasks including information collection at the time of the emergency, and they are equipped with sensors for recognizing the environment. Auditory information is considered in this paper since it can be used in cooperation with other information such as vision in order to enhance environment recognition. Although MUSIC is the robust signal processing algorithm that is commonly used for sound source localization, the noise generated by rotating rotors that is called ego-noise distorts the target signal significantly, which leads inaccurate estimation of the target source direction. This paper proposes a method to attenuate inaccurate estimation triggered by ego-noise.

Our research goal is to improve a status display of information acquisition system for unmanned construction machines to conduct an initial response (within 1 month) of sediment disasters. One of our approaches is to fly a tethered multi-rotor UAV (unmanned aerial vehicle) that mounts a camera, and to obtain environmental information around the construction machine. In order to advance this idea, we conducted a simple flight test of our multi-rotor UAV around a constructed machine. In this paper, we report the initial test, and introduce our lessons learned.

This paper proposes a task space P-SD servo controller with the finger-tip tracing motion. This controller consists, joint-space PID controller, task space velocity feedback with saturate function and visual feedback controller. Details of the proposed controller and the satisfaction of the passivity were mentioned. For achievement of the task space velocity feedback, a tactile sensor was developed by the knowledge of the human skin-structure. The sensor consist an epidermal-dermal dynamics, and PVDF velocity sensors. Some experiments were conducted to verify the validity of the proposed method.

This paper proposes a control strategy of a tendon driven mechanism based on the analysis of human like motion. Based on the human like muscle activities, proposed controller uses both the discrete-time and the continuous-time controller. The continuous and the discrete controller activate the muscle to generate the force fields. Especially, the discrete-time controller changes the number of the active muscle. The controller can select the number of the muscle units.

This paper describes a new method of operating the welfare support equipment with gazing input. User's face was observed by the USB camera, and it was taken to the control computer to recognize the user's gazing directions. Gazing points with respect to right or left directions could be recognized by the relative distance of an iris and the outer corner of left and right eyes. Upward and downward gazing direction were judged by the curvature of the eyelid shape which was approximated by the Bezier curve. As the gazing point became lower than the height of the camera, the curvature value of the approximation curve was decreased because the eyelid shape roughly became linear. In the test trial, it could be confirmed that the developed system could recognize the user's gazing directions with non-contact state.

Many mathematical models have been proposed for Vestibulo-Ocular Reflex (VOR) of human from head movement. It is known that, the model accuracy of these conventional models decrease when mental task is added to the human. On the other hand, 6DOF-SVC model, a mathematical model of motion sickness due to vestibular sensory system, includes a sub-model block representing the prediction accuracy of their own body movement, which is thought to be affected by the mental task. Therefore, toward description of the influence of mental task on VOR, the proposed VOR model connects the estimated single of body movement in 6DOF-SVC model to the VOR calculation path. The parameter identification experiments were coducted which to compare the accuracy of the proposed model with that of the conventional model. We found that the accuracy of the proposed model is as high enough to describe human VOR as the conventional model.

It is considered that gazing motion is very useful due to the fact that the motion is unburdened and non-contact operation can be done. In this paper, we proposed a new intuitive interface based on the eyelid shape. User's face was observed by the USB camera, and it was taken to the computer to recognize the user's gazing directions. The gazing directions in the upward and downward were recognized by approximating the eyelid shape with the Bezier curve. As the gazing point became lower than the height of the camera, the curvature value of the approximation curve was decreased because the eyelid shape roughly became linear.

This study proposes a novel approach to the FBI study, in which experimental participants can voluntarily interact with their own body (virtual own body) in a second-person point of view, to reveal the underlying mechanism of human body ownership. The present paper mainly explained the proposed experimental scheme and described its novel aspects to the FBI study. Additionally, an experimental system, which employed a robotic master-slave system and a virtual reality, was designed and developed to enable the self-interaction. In this study, we confirmed the applicability of the experimental system to the proposed FBI study.

AIKIDO is a famous Japanese martial art and contains some interesting movements because a master of AIKIDO can control others only by small power. In this research, we measured motions of ”Tenkan Movement” which is one of basic technique, by motion capture system and surface myoelectric potential sensors, and discuss about relative positional relations between the AIKIDO practitioner and the opponent.

In recent years, there has been a growing needs of the sensing system to acquire the position and attitude of the people in the field in the medical and sports. As future, it will be possible to quantitatively determine the individual differences and proficiency in body motion. The purpose of this paper is to perform feature extraction using Singular Value Decomposition for human gait data.

This paper suggests a new method of recognizing a facial orientation based on tracking a pair of nostrils. Due to the fact that nostril shape is stable compared with other facial parts such as eyes or mouth, the nostril image was treated as a feature point and it was taken in real-time for image processing via an USB camera. When a man faced to up and downward, the darkness areas of both nostrils were increased and decreased, respectively. On the other hand, the difference between both nostril areas could be caused in cases where the face was turn to the left or right directions. The obtained characteristics were utilized for recognition of a facial orientation.

We developed a novel tongue-operated joystick device for severely disabled persons. This device has a function to control the reaction force against the tongue operation. The user can operate skillfully a powered wheelchair or a robotic arm by using force feedback from this device. Through the change in reaction force, the device can notify the user whether the interference between the equipment and the environment occurs. In this paper, we introduce the detail mechanism of this device and examine whether the user can perceive the change in the reaction force.

Power assist devices are very useful for reducing the burden of, especially, elderly and women operators who involve in hard work tasks. In the operation of the power assist device for carrying an object, the operator not only has a feeling of wrongness but also does dangerous operation if there is a big deference between an expected and sensed feeling of the weight. The feeling deference can be reduced if the sensed feeling of the weight can be controlled without changing the burden to the operator. In this report, a method for controlling the sensed feeling of the operator and the effectiveness is shown experimentally.

In this study, we measure the body movement of a car driver by using three-axis acceleration sensors for detecting an aimless driving state. The data of driver's body motion in a driving simulator is collected from 10 subjects. These results suggest that the aimless state can be detected from the variance data of the acceleration on the right wrist.

Human beings gain more than 80% of information through the eyes. It is noted, therefore, that line of sight information provides an important clue to various information, including psychological information, such as human attention or interest. This paper focuses on saccades, which are considered to be the fastest eye movement. It was generally thought that perception during saccades is blocked. More recently, however, it has been suggested that perception occurs during saccades The purpose of this research is to develop a fast and accurate gaze tracking system that can be measured saccade. Although the limbus tracking method shows a high-speed response, it is not accurate enough. Therefore, we propose a high-speed and high-accuracy gaze tracking method with multiple light-emitting and light-receiving elements.

Recently, Brain-Machine Interfaces has attracted attentions for many researchers. Most of these BMI researches try to support the disabilities. The purpose of this study is to develop a new BMI to better assist and support the disabled person's life. In this paper, we designed experiments to investigate changes in EEG while people with observes or imaging the movement based on the Mirror Neuron System.

In this study, we aim for clarifying a way to improve driving skills in motorsport. Perception of vehicle state is very important for driving, especially in motorsport. In this paper, therefore, we first observed driving behaviors of a skilled driver and a beginner using a driving simulator and compared their steering behaviors and eyesight. Based on the comparison, we investigated effects of gaze point on steering. The results show that an appropriate gaze point was neither near nor far from a driver, it was in between them. Finally, we demonstrated the effectiveness of an instruction of gaze point for improving driving skills of beginners.

This study aims to develop of wearable sensor system, and propose an estimation method of lumbosacral alignment and lumbar load caused by changes in posture. Need to use the hand load for lumbar load estimation. Therefore, the hand load estimation method is developed by using muscle hardness sensor. By using this method, the hand load can be estimated with enough accuracy.

In this paper, we construct 6 types of multi-class classifiers and compare the performance of them. In the feature extraction phase, CSP filter (common spatial pattern filter) and PCA (principle component analysis) and LDA (linear discriminant analysis) are used. In the classification phase, kNN (k-nearest neighbor) method and SVM (support vector machine) are used. As a result, the best classifier for discriminating grasping patterns was CSP/kNN classifier (We set the selective number k as 1). The best classification rate for 9-class power-type grasping patterns of CSP/kNN classifier was 48% and the classification rate for 7-class precision-type grasping patterns of CSP/kNN classifier was 40%.

In the nearby special support school, driving practice of motorized wheelchair intended for the physically disabled child have been conducted. We developed the joystick-type controller that adjusts the reaction angle automatically to suit the actual condition of the physically disabled child so far and it is lead to practice the driving of the motorized wheelchair. In this paper, aimed to examine the effectiveness of the proposed method by analyzing the gaze behavior in a driving of the motorized wheelchair and proposed method.

Recently, the fatal accident of elderly people has increasing in Japan. A wearable assist machine for elderly people is studied as one of countermeasure. But, the size of assist machine grows big and the weight of it gets heavy if its degree of freedom increases so as to prevent obstructing the movement of human. The purpose in this study is to extract the major modes of sit-to-stand movement with restraint of human trunk for design small and lightweight assist machine by fixing degree of freedom on the optimal position. Measurement of sit-to-stand movement was used motion capture system. The major modes were extracted by the mode analysis. As the result of analysis, the coxa and knee flexural extension, adduction and abduction ware identified as major modes of human sit-to-stand movement. These motions were the movement to lift a body and opening and shutting movement of the leg with standing-up.

This paper proposes a measurement system for simultaneous measuring brain waves of human and cenine. The system consists of a headset for human brain wave, a headset for dog brain wave and video camera. The headset for a dog brain wave measurement is specially developed by modifying human headset. The fabric cover attached with electrodes is fixed to a dog's head with Velcro. The ears of a dog can extend naturally from holes of the fabric cover, which made a longer time experiment possible. The proposed system does not need any cable: the measured signals are sent to a computer through Bluetooth, and the electric power is provided by battery. The experimental test is conducted with a police dog.

We found when a people decide their action based on visual information its reaction time can be shorten by adding tactile stimulus. Our experimental result shows that the reaction time can be shorten more than 10 % with tactile stimulus than without giving it. This phenomenon occurs not only the situation without judging but also that requires judgement which direction to move. Our experimental result also shows that the tactile stimulus method does not affect to the result.

In order to establish the method estimating the stability of human walking, we propose evaluation of the stability based on ZMP theory through a simulation of human walking. According to the result, we confirm that ZMP is located within support polygon when the walking is stable and ZMP is located out of support polygon when the walking is unstable. Therefore, we could say that it is correct to estimate the stability of human walking based on the ZMP theory.

This paper examines a relationship between an arterial stiffness parameter β and nociceptive mechanical stimuli. The stiffness parameter β which reacts according to peripheral sympathetic nerve activity is calculated from continuous arterial pressures and photoplethysmograms based on the log-linearized peripheral arterial viscoelastic model, while mechanical stimuli were applied to the subjects in the experiments. The results showed a significant correlation between the stimulus velocity and the stiffness parameter β when stimulus amplitude was constant (r = 0.63 (p < 0.05)) and that between the stimulus amplitude and the normalized stiffness parameter β_n when stimulus velocity was constant (r = 0.54 (p < 0.01)). It was therefore concluded that the arterial stiffness parameter β may react to amplitude and velocity of mechanical stimuli.

For teaching the skill of brushstroke in calligraphy, a teacher often holds both a hand of student and an ink brush, and moves them to appropriate direction. Since no scientific evidence about the effect of this teaching method has been reported, then this study discusses the mechanisms of improving the proficiency level of human skill.

There is a tendency that the elderly persons decline in feeding and swallowing. It is known that the oral training such as a facial exercise has an effect to facilitate a smoothness swallowing motion. This paper describes measurements of face exercises using PET plastic bottle and evaluations based on the electromyography (EMG). Here, the orbicularis oris muscle were focused for an evaluation. As results of APDF analysis, it could be confirmed that the motion of keep exhaling and fast exhaling, and using thin mouthpieces had effects to enhance the activity of the orbicularis oris muscle. Therefore, this paper also describes a training system based on results of analysis.

This paper describes the human face detection based on special feature color. Here, the skin color was picked up as a special feature. The human's face was observed by the USB camera in real time. First of all, human face images were obtained under the various brightness environments. As a results, it was found that red component of the facial skin is stronger than any other facial parts such as eye, nostril or mouth. In our proposed system, the original image was changed into the rough color image and the smoothing procedure was applied to calculate the weight value of the skin color and eliminate noise. In cases where the head was tilted or the face was turned to the side, it is difficult for the existed method to detect the face, because it was adjusted for the frontal face. On the other hand, our proposed system could detect the tilted face and side face.

In this paper, we propose a method for estimating the contact force of fingertip by inputting vibration actively. Usage of active bone-conducted sound sensing has been limited to estimating the joint angle of elbow and finger in our previous researches. Therefore, we applied it to the method for estimating the contact force of fingertip as another application. Unlike conventional methods, it is not necessary to mount the device on a fingertip, and haptic feedback is also enabled.

EEG signal is recorded from the human scalp non-invasively and can be useful to control wearable robotic devices, according to the human motion intention. However, due to its high information density, the estimation of the user's motion intention from EEG signals is not easy. The user's motion intention might not be estimated when the user does not concentrate on the control of the robot and is distracted by other things. In this paper, a neural network based real-time estimation method is proposed to detect human motion intention using EEG signals. Results show that EEG signals can be successfully used to predict the 2-DOF motion intention of the user.

In recent years, the news of the wearable assist device suitable for work in the farming and factories has become a hot topic. In this study, we propose a new spring mechanism named as Two-wire-three-mode spring mechanism that has three phases of stiffness properties. We developped a experimental equipment using Two-Wire-Three-Mode SAT for rotating joints for verification its basic mechanism. Based on experimental results we confirmed that the mechanism has automatically changed the joint stiffness from compliant to stiff progressively by three stages according to the rotation angle of the joint. We also applied the spring device for a wearable posture assist device which has no power devices. We developped a prototype model that consist of link along the body surface from the back to the ankle, the spring device, ancker unit and wires.

We proposed a stretchable and bendable electronic device by folding deformation. We used Miura-ori that is a kind of origami folding for folding deformation. However, original Miura-ori is a rigid folding and its deformation is restricted. When Miura-ori bend to out-of-plane, it has saddle-shaped deformation mode and it is unable to stick on sphere. Therefore, we developed Miura-ori which has holes on intersection points of mount and valley creases to be enable to bend without saddle-shaped deformation and stick on sphere. In this paper, we evaluated effect of hole size on the interaction points that is enable to stick on sphere and concluded 30% of one side of the rhombus plane in Miura-ori is appropriate. As a result, we achieved an electronic device which is possible to folding, deployment and bending deformations without wiring cracking.

The study related to the thermoregulation of exercise under hot environment has become an important issue from the point of view of preventing heat disorders and improving performance. Recently, there are many studies about thermoregulation under exercise. However, measurement of their study has a limit on the number of sports, and cannot monitoring of tympanic temperature change in real time. Consequently, there is no practical measuring device. In this study, we will develop the wireless tympanic thermometer that has not a limit on the number of sports and can get the trend of the tympanic temperature change in various exercises. Furthermore, we will measure tympanic temperature in practical exercise, and aim for practical application.

This paper shows effect of parallel springs with a leg during the stance phase of running. Effect of series springs on human running has been already revealed. On the other hand, the effect of parallel spring is not clear because of complex mechanism of device: the device locks the spring to store kinematic energy during the initial stance phase and to restore the energy during the terminal stance phase, and the spring is released during the swing phase to allow free leg motion. The authors'device consists of sliders, leaf springs and continuous lock mechanism. A runner wears this device on the waist and the heel, and the leaf spring supplies leg stiffness. Some experiments reveal the effect of the leaf spring.

We have proposed a robot suit with two hardware-based safety devices - the “velocity-based safety device” and the “torque-based safety device”. The robot suit assists a patient's knee joint. The torque-based safety device stops the robot suit if it detects an unexpected high joint torque. Similarly, the velocity-based safety device stops the robot suit if it detects an unexpected high joint angular velocity. The safety devices work even when the robot suit's computer breaks down, because they consist of only passive mechanical components without actuators, controllers, or batteries. The transient responses of the velocity-based safety device to some ramp inputs have been analyzed. In this paper, we verify the effectiveness of the analysis of the velocity-based safety device by experiments.

In our previous work, a small wearable robot named telecommunicator has been developed. The telecommunicator can achieve rich communications with a remote person. Until now, we developed a prototype of the wearable miniature humanoid MH-2 as telecommunicator. It has a 3-DOF head, a 3-DOF body and 7-DOF dual arms. MH-2 also has human like shape and ability to reproduce the movements of human's arm and head. By using a real avatar, we are aiming to achieve a high fidelity communication system. However, the current head mechanism reduce reality because of its inaccuracy. In this paper, a new 3-DOF head mechanism is introduced and designed.

Electromyogram (EMG) signals represent the electric potentials on the surfaces of muscles. These signals may be applied for controlling prosthetic replacements such as a prosthetic hand. This paper treats on application of EMG prosthetic hand for congenital defects. The experimental product of EMG prosthetic hand is developed for the subjects under 5 years old. The main problem of experimental product is a design issue of socket of amputee. The socket needs to support all weight of prosthetic hand and object of handling. Furthermore, EMG sensor electrodes are needed to be attached on the surface of amputee side firmly, in order to avoid noise. Therefore this paper proposes new socket design by using conductive polymer and flexible fabric. The experimental result shows that the prototype socket is applied for infant amputees and succeeded to grasping and releasing small object and holding big heavy object by using EMG prosthetic hand.

We have been developing the wearable robot “muscle suit” for direct and physical motion supports. The use of the McKibben artificial muscle has opened the way to the introduction of the muscle suit - compact, lightweight, reliable, wearable “assist-robots” enabling manual worker to lift and carry weights. However, the necessity of compressed air for the McKibben artificial muscle set a limit on the application field. So we developed the muscle suit which does not need any external air supply.

Entertainment is one of the successful fields in robot industry, and especially, the typical entertainment robot is the small humanoid robot. To offer the people new elements of entertainment, we focused on figure skating. In the previous research, we realized a straight sliding and turning motion for small skating robot. However, the motion is small and not attractive. So, we want to realize more dynamic motion; however, it caused robot to be unstable. In this study, we discuss more dynamic skating motion and the why to stabilize the robot.

In this paper, we propose an algorithm for an air-hockey robot to decide its optimal motion based on data of human responses. The trajectories of the puck in a lot of games by humans are recorded, and they are approximated by Gaussian Mixture Models. The model is used to predict human motions, and it is integrated into a 2-step look ahead algorithm. The proposed method is verified by numerical simulation.

A sound pitch online feedback algorithm is proposed to adjust a correct sound of a violin playing robot. A relation between a finger position pressing on a bowstring and a sound peak frequency detected by FFT processing can be expressed by a quadratic equation. The amount of movement of the linear actuator which moves the finger is calculated based on the quadratic equation from the difference between the target pitch and the present pitch, and the linear actuator moves. By repeating these operation, the correct pitch is adjusted. A sound pitch adjustment experiment is examined using the proposed online feedback algorithm, and we confirm that the convergence of the online feedback operation is about 3 times.

The ultimate goal of this study is to develop an anthropomorphic violin-playing robot that can expressively perform music pieces. In this study, we investigate how violinists alter sound volume and tempo in two musical pieces to express “bright” and “dark” timbre. Two trained violinists participated the experiment. We analyzed the data and found that the sound volume of the bright performance is larger than that of the dark. We also found that the sound volume of the dark performance increased in the middle of the musical piece. The violinists took faster tempo in the bright than the dark. In addition, we designed sound volume which varies with time based on the obtained results.

This paper deals with development of a left hand finger for our violin-playing robot with two anthropomorphic arms. Our previous left hand uses RC servomotors, which leads large mass and prevents the left arm from its vibrato motion. Therefore the left hand should be lightweight as much as possible and we adopt wire-driven mechanism for our new fingers to place actuators to the outside of the robot. We manufactured a finger with two joints: MP and PIP. It is necessary to be able to play the violin. We confirmed that the developed finger has enough moving ranges. Also we conducted an experiment and confirmed that it produces enough force needed to push the strings.

Entertainment is one of the successful fields in robot industry, and especially, the typical entertainment robot is the small humanoid robot. To provide new elements of entertainment for the people, we focused on figure skating motion. In this study, a software which choreographs any motion to any robot and generates robot's motions was developed. The software was developed as a plug-in for “V-REP” (Virtual Robot Experimentation Platform). The user can easily make key-poses on the software, and then, interpolation motions between key-poses are automatically generated. Its usability was confirmed by a user test.

This study aims to develop an intelligent bed which can control body pressure distribution adaptively. This bed is actuated by multi-flexible bags. This bed is controlled by air pressure conjugated with a body simulator. This report describes basic experiments with multi-flexible bags and a fundamental body simulator. These bags realized high pressure resistance and we verified experiment of feedback control by using inner pressure sensors. For development of the body simulator, we examined a simple dynamic model to obtain body pressure distribution and body posture in lying position.

In this report, we propose an operating support system that leads the operator to a next estimated operation by a guide force. If operator feels that the estimated operation is different from his, he can take his. The estimated next operation is selected to compare the current operating sequence and the models integrated the past operating sequences in the same situation. We confirmed that the proposed system was efficient to drive an actual small vehicle with and without guide force.

Human beings are doing a smooth communication with the auditory, visual and haptic information. Multimodal interfaces are utilized the communication operation assist could be improved by visual and auditory information given the time difference in presentation. In our previous research, the validity of multimodal assists with time difference were experimented by vibration control task for a pendulum. The experimental results showed that giving a time difference to the information is effective in the machine operation performance. However, the results depended on the task. Therefore, the experiment is conducted by drive simulation, as a more general task. And in this paper, validity of proposed assist with the tasks is verified, as a first step, the effect on the operating characteristics of each assist. Finally, experimental results show that giving the visual assist is effective in reducing the angle error, and the auditory assist is effective in reducing the position and angle error.

Recently, the neural network is used in teamwork assist studies to predict the solo task model. Prediction accuracy of the solo task model is required to be investigated. Therefore, prediction accuracy of the solo task model was verified with general joysticks. However, prediction accuracy of neural network has not been verified when predict the solo task models with force feedback joysticks. Furthermore, the study uses a driving task. However, in this study, oar model force feedback joysticks are developed which used the rowing boat task in the future. Accordingly, in this study, we verify the prediction accuracy in the neural network learning with oar model force feedback joysticks.