Force measurement plays an important role in securing safety in the nursing care field due to the requirement of human contact and interaction such as lifting up human. Particularly, it is possible to perform delicate operation to the human by detecting a minute force such as delicate balance and contact of a human. Conventionally, we have realized a load sensor using quartz crystal resonator (QCR) with a measurement range of the order of 10⁶. However, the conventional QCR load sensors were limited only to the load measurement. In this paper, 3 axis force sensor using QCR load sensor was proposed in order to expand the application of the QCR load sensor. We also conducted the calibration of the force characteristic of the fabricated QCR force sensor.

This paper proposes a wearable device mounted on a human fingertip which measures the 3-axis force on the fingertip (both normal and shear forces) exerted while grasping or touching. The device consists of two Hall-Effect sensors attached on both sides of the fingertip to measure the finger pad deformation caused by external forces on the palmar side of the finger. The experiments in this paper proved that the device can measure both normal force and shear force. The recorded forces could be used for example to record the forces that human exert while manipulating objects.

Movements include personal characteristics and emotions. This study aims at estimation of personal characteristics and emotions by using tactile information. The tactile information is private and its applications could be available for various things like security and communication. Especially, we focus on skin vibration as the tactile information since our tactile sensations are based on our skin deformation and vibration elicited by the mechanical interaction between our skin and an object. As the first step, this paper investigates personal identification using skin vibration at key touching. We use the wearable skin vibration sensor that measures skin-propagated vibration and allows users to touch an object with the bare finger keeping natural tactile sensation. Skin vibrations are collected during 10 days for 10 participants and personal identification is conducted with deep learning. Results demonstrate that skin vibration can be available for personal identification.

Walking experiences are affected not only through the visual sensory but also through the walking sensation. There have been few studies focusing on the relationship between a person's subjective feeling while walking and physical information he or she obtains from the ground through walking actions. This research investigates the relationship between the subjective feeling while walking and the physical interaction data with the ground by using the measurement device we have designed to be mounted on regular shoes.

This paper shows a method for achieving multi-contact humanoid Motion On a Movable Object (MOMO) such as climbing of a stepladder. MOMO has the following features: various constraints need to be considered and it has to handle scenarios where the mass properties of an object are unknown. In this paper, in order to achieve a humanoid robot having MOMO, we propose balance constraints including object-environment constraints and an online estimation of object's constraints. We evaluate our proposed method through experiments in which a life-sized humanoid robot climbs stepladders that have unknown mass properties.

This paper presents the development of a simple hydraulic robot for experimental evaluation of various dynamic 3D balance controllers. The robot has three links including the foot plate, which are connected by two universal joints. The each joint is actuated by two force-controllable hydraulic servo actuators having the coupled configurations. We show the initial dynamic simulation results using a LQR controller to verify the fabricated robot has enough potential for balance control experiments.

In undeveloped environment, it is necessary for humanoids to move from high place to low place without stairs or ladders. We propose rappelling applying belt friction by humanoids as a method of vertical movement. Rappelling motion we proposed consists of two parts, generation of legs' pose for rappelling and reliable rope manipulation.

This paper aims at a novel framework of evaluating supportive effects during dynamic motions supported by wearable assistive devices, by extending our previous work that has proposed a humanoid-based method for evaluating static torques. For this purpose, first we introduce a model of the human intension of motions supported by active devices by observing respective human tasks. Next, evaluation of assistive devices during dynamic motions is addressed by a torque controller for a humanoid that tracks retargeted human motions. Finally, the effectiveness of proposed method is validated by experimentally assessing assistive device ”Muscle Suit” actuated by pneumatic artificial muscles.

The human forearm is composed of two long, thin bones called the radius and the ulna, and rotates using two axle joints. We aimed to develop a forearm based on the body proportion, weight ratio, muscle arrangement, and joint performance of the human body in order to bring out its benefits. For this, we need to miniaturize the muscle modules. To approach this task, we arranged two muscle motors inside one muscle module, and used the space effectively by utilizing common parts. In addition, we enabled the muscle module to also be used as the bone structure. Moreover, we used miniature motors and developed a way to dissipate the motor heat to the bone structure. Through these approaches, we succeeded in developing a forearm with a radioulnar joint based on the body proportion, weight ratio, muscle arrangement, and joint performance of the human body, while keeping maintainability and reliability. Also, we performed some motions such as soldering, opening a book, and badminton swinging using the benefits of the radioulnar structure, which have not been discussed before, and verified that Kengoro can realize skillful motions using the radioulnar joint like a human.

In a previous paper, we presented a control method for master-slave bipedal walking robots. The method consisted of a combination of position-based unilateral masters-slave control of leg joints and balancing control actuating ankle joints. This paper presents an improved switching mode method and balance control laws assosiated to the modes. The proposed method is validated through a realtime simulation environment involving haptic devices.

Humanoid robots should be able to keep balance for any direction of disturbance. N-step capturability framework was proposed for fall avoidance that involves a number of steps. But it did not consider collision between the support leg and the swing leg in its analysis. For this reason, the conventional capturability analysis cannot be applied for a case in which a robot is pushed in sideways. In this paper, we analyze sideways legged locomotion in terms of capturability. Based on this capturability, we propose the sideways legged locomotion for fall avoidance. The proposed method is validated through numerical simulation. The simulated robot was able to recover from predetermined sideways pushes. It was also found that robots can be able to recover from larger disturbance with variable step length.

In daily-life environments, humanoid robots are expected to perform household tasks instead of humans. Humanoid robots can move over a large workspace but generating their motions is still very time consuming. We aim to improve the realtimeness of whole-body motion generation for humanoid robots. In this paper, we introduce a rapid whole-body controller which uses regression of the torso posture. By estimating the torso posture, we reduce the calculation cost of inverse kinematics, as we solve it separately for the legs and arms. We apply the proposed method to the humanoid robot HRP-4 for the task of reaching low-height positions which requires hunkering down. This motion takes 3 seconds.

In this paper, we developed a bipedal robot and experimentally verified balance control based on ankle angular momentum. This balance control maximizes stabilizable regions and can optimally switches balance controllers. These balance controllers recover balance by mainly using ankle torque, inclining trunk posture and stepping respectively. In contrast with balance control that is based on position and velocity of center of mass and ignores robot rotation around center of mass, this balance control works well even when the robot rotates around the center of mass. Experimental results demonstrated that this balance control recovered balance of the developed robot by appropriately switching the balance controllers that maximize stabilizable regions.

The KBHR (Kanagawa Biped Humanoid Robot) that is composed of 42 DOFs have been developed. This paper describes a trajectory generation method for 7-DOF manipulators of the KBHR and a robot vision-based object recognition method. The trajectory generation method is based on numerical analysis using Jacobian pseudo-inverse and transpose matrix. In order to recognize objects, an OpenCV image processing library is used. Through an experiment recognizing an object and handing over it to a person, the effectiveness of the trajectory generation algorithm and the object recognition methods is confirmed.

In this paper, I describe turning walk on loose soil of a humanoid robot. I propose landing position control to curb ZMP deviation on loose soil. I inspected effectiveness which is use a miniature humanoid robot.

Although how to perform tasks at disaster sites has been actively discussed, there is little debate on important issues such as how to deliver robots to disaster sites. We propose to drop a humanoid robot with a parachute from an airplane as one of the methods of delivering robots. In this research, we aim to establish a method to absorb the shock at the time of landing by whole body motion for preventing the damage to the robot. As an early stage of this research, drop tests using a small one-legged robot were conducted. From the viewpoint obtained by these experiments, we designed a parachute landing motion for the small robot heuristically and succeeded in reducing the impact force to 40%.

Recent years, worker shortage in the field of urban construction is becoming a bigger issue. To solve this problem, direct-vision remote control of a construction machine can be an effective solution where a worker can play not only the role of an operator but also the role of manual laborer. Authors propose a new direct-vision remote control system for an excavator using a tablet device. In the system, the graphical user interface (GUI) enables an operator to control an excavator by using various multiple-fingers action. An excavator teleoperation system has been developed to conduct an experiment. In the usability-testing experiment, the usability of the proposed system has been evaluated and its performance has been compared with a GUI that simulates an interface of conventional machine levers. The experimental results have verified the advantage of the proposed system.

Techniques using drone as a next generation infrastructure inspection have been proposed. However, it is difficult to stably flight control to approximate drone to inspected object. In this research, we aimed for more accurate flight control by obtaining the relative distance and the relative angle around yaw and pitch for the drone to the wall surface by using the laser range sensor.
As a result, by using information on the wall surface obtained from the sensor, accurate flight control of the wall surface scanning which is difficult by human maneuvering can be performed.

In this paper, we report on an inspection system integrating vertically moving winches and a prototype inspection carrier in the development of the robot system that inspects the inner wall of a large boiler for power generation. We develop a mechanism that can raise and lower while maintaining the horizontal level of an inspection carrier using two vertically moving winches and integrate them with an inspection carrier. We will install inspection equipment on the inspection carrier for inspection operation and construct an inspection system. And we are planning experiments using a large imitation wall to evaluate the performance of the prototype inspection system we developed, and from now on we will confirm the behavior of an inspection carrier and identify problems.

In order to realize visualization of stress acting on infrastructures, the authors developed “Strain Visualization Sheet” based on the principle of Moiré fringes. Strain Visualization Sheet is a revolutionary device which makes it possible to check strain values with the naked eye, and also enables highly precise measurements of strain values by using an ordinary digital camera. As an additional advantage, it is not necessary to provide an electrical power source for the device. In addition, Strain Visualization Sheet is a self-temperature compensated device, it is unaffected by temperature during strain measurements of the concrete and steel materials used as construction materials in infrastructures, and strain measurement with an accuracy of > 20 με is possible. As Strain Visualization Sheet is low in cost and enables any user to grasp the stress acting on infrastructures in a simple manner, we are confident that it can make an important contribution to operation and maintenance (O&M) of infrastructures in the future.

Recently, several types of wall climbing robots have been developed for the maintenance task of building, bridge and so on. A leg-type wall climbing robot is one of the solution for such tasks. In this paper, we propose the progressive wave creation mechanism to improve an wall climbing ability. Through the several experiments, the effectiveness of proposed mechanism are confirmed.

Hammering test is a widely used for non destructive inspection of concrete structure. It is important for reducing human's burden and time-consumption because there are large and a lot of concrete surfaces which should be inspected. This study introduces development of hammering test support system by displaying the diagnostic results of hammering test via projection mapping. This system detects the hitting positions, analyzes the acoustic data of the hitting sounds, and shows both results on the concrete surface in realtime. We performed the evaluation of the proposed system through the measurement of the accuracy of the hitting positions and its defection on test concrete block. Moreover, we evaluated the proposed displaying method compared with non-support approaches.Through the experiments, we confirmed the effectiveness of the proposed system.

Four-legged animals achieve high speed and high efficiency movement by changing their gait according to speed and terrain. Consequently, it can be assumed that such movement is possible by a four-legged running robot based on reproducing the various gaits of four-legged animals. We are developing an active-suspension mechanism consisting of a spring and a DC motor on the leg module to imitate the legs of animals and studying the high-speed running of a four-legged running robot. This mechanism strengthens the instantaneous power of jumping and controls dumping at landing, although the springs and motors are mutually constrained and it is difficult for them to demonstrate maximum force. In this study, we examined the optimal spring for an active suspension leg module. Furthermore, we evaluated the running speed and power consumption of a trot and a bound gaits using the selected spring.

We aim to make a one-legged hopping robot hop continually by employing a neural network-based controller, which is trained through remote balancing control by humans. Consequently, the height and stability of the hop become important for human control of the robot's balance. Our robot uses an active-suspension mechanism. The robot performs dumping with motors at foot landing and exerts the instantaneous power of a spring to jump. However, it is difficult to achieve a stable jump with sufficient height because the spring and the motor are mutually constrained. Therefore, we introduced a mechanical or electromagnetic clutch mechanism into the robot, and succeeded in increasing power of the jump. In this study, we re-examined the spring and the motor in terms of their abilities to improve the quality of hopping. In addition, we introduced a new control that reduced the preparation time for the next jump and restrained rebounding at landing.

We developed a hopping robot using two oblique inertial rotors. The robot has 4 DOF, and consists of a body, one leg and two oblique inertial rotors. The robot can move three dimensional space using the inertial rotors. This paper shows the realization walking control using two inertial rotors. We conducted an experiment to simultaneously perform walking and balance maintenance. The results and considerations are described.

Although legged robots have been researched for more than 20 years, their movement is not quick as same as animals. Abilities of a leg are not clarified. The objective of this research is to clarify the abilities and functions by manufacturing a one-legged robot. The one-legged robot is designed as almost same as human legs. Since it has not only electric motors but air cylinders in its thigh and calf as actuators, it has enough power to move. Local PID controllers for body attitude and center of gravity work to stabilization of standing control for the one-legged robot.

The continuous motion composed of rolling and jumping such as continuous handspring is useful as the locomotion of robots. In order to realize the continuous handspring, we propose a multi-link robot model composed of two limbs and a body and the control law for the model. We verified the controller using the combination of attack angle control and the control based on the angular momentum of the robot in the simulation. In the verification, the robot model realized 100,000 steps in the continuous handspring in the certain environment. In addition, it realized 20 steps after the temporary change of coefficient of static friction of the floor from infinity to four and over in the verification of the robustness against the temporary disturbance.

This study aims at proposing a mobile robot that can move forward on snow stably. For this purpose, this study will show the designs of a leg which suit to move on snow. Legs have advantages of stepping over hard obstacles, which is good for moving around rugged road. However, they could be buried in snow by pressure because the grounding areas are small. To solve this problem, we pick on a snow-shoe. Snow-shoe is the device which fit under the shoes when walk on snow for human. There is no study of using snow-shoe for robot, so we propose the snow-shoe which is automatically folded while in swing leg phase.

Maintenance of infrastructures is a major problem. In order to streamline it, automation of inspections is researched. For complete automation, the measuring position determination method of inspection machines is necessary. This paper presents a method of measuring position determination using vector quantization for automatic visual inspection systems using Unmanned Aerial Vehicles (UAVs). In the proposed method, it is assumed that the 3D model of the target infrastructure, the specification of measurement and UAV’s features are given. Under those assumptions, the 3D model is regarded as the cluster of vectors and the measuring positions are determined using vector quantization. In this paper, the proposed method is compared with the conventional method by a computer simulation assuming visual inspections.

This paper describes about the driving control of a quadrotor propulsion wheeled mobile robot for the purpose of the application to the inspection of the social infrastructures. This quadrotor propulsion wheeled mobile robot can move the ground and the wall surface and the ceiling. The optimal servo control system was adopted in order to control the turning motion. Comparing the simulation and the experiment, The tracking performance of the system was confirmed for the target turning angle in this paper. On the other hand, So that a robot moves along a wall surface, it is required the measuring the angle of the robot to the surface of the wall in order to face the wall. The robot was equipped with two infrared range sensors. From experimental results, we confirmed the effectiveness of the angle detection with the optimal servo control system and the infrared range sensor.

Recently, there has been an increase in the number of maintenance inspection management with the rapid aging of social infrastructure. Image-processing technology applied for inspection to as a more efficient maintenance. Above all, to construct inspection record takes times and cost. Therefore, it is necessary to construct automatically damage site map. In this paper, we proposed the method of automatic construction for inspection of concrete structure by image processing using an affine transformation and laser pointer. We create a damage site map that showing the damage position of the crack. It was possible to improve the efficiency of inspection work with the automatic construct method of damage site map.

In the purpose of the development of an automated system to detect cracks and internal holes in the concrete infrastructure, a mechanical system to reproduce hammering tests is highly important. The system is required to provide an effective impact force for the accuracy in acoustic analysis. In the present study, we focused on bifurcation phenomenon of the nonlinear spring system in the form of the buckling instability, proposed a prototype of the hammering system based on the spring dynamics to provide a quick motion of the hammer head and analyzed the energy theoretically.

To improve the inspection method in the inner wall of the tunnel that became superannuated, the mechanism that was flexibly transmutable of the guide frame shape for the inspection according to the object in the tunnel was developed. In part 2 of the report, to decide complex frame shape to avoid several objects on the tunnel ceiling side, the technique for simulating actual frame shape was examined analyzing the shape of the guide frame by spline interpolation as mathematical technique. The shape of the guide frame can be simply analyzed by generalizing this technique. In this paper, overall plan to evade objects, mathematical technique for shape analysis, those analytical result and the utility are examined in details

To improve the inspection method in the inner wall of the tunnel that became superannuated, the mechanism that was flexibly transmutable of the guide frame shape for the inspection according to the object in the tunnel was developed. In part 3 of the report, three dimension laser range finder of the high-resolution evaluation was constructed, and the position of the obstacle in the tunnel was detected The center and the depth position of the obstacle were presumed from the relation between the distance and the angle, and the boundary condition that changed the shape of the guide frame was calculated. The shape of the guide frame was analyzed based on the location information of the obtained objects.

Carrying robots to work place where got disaster sites is dangerous. If they can be thrown into the place, operators will be safe. The purpose of this study is development of durable jumping robots. Thus, we have been developing a robot with spherical shell. This robot has three legs and function of jumping to any direction. In this paper, we report consideration of two walking motions of three legs. Walking motions are adopted to “Creeping-type inchworm gait” which is using body. Finally, a result of the experiment of these is shown.

For autonomous hexapedal locomotion of a legged robot, SHIKYAKU, we developed a teaching playback system. Since the mechanism of SHIKYAKU is based on the extension and flexion of fingers, motion patterns of the robot for walking and climbing are taught with finger through LeapMotion. The robot records the motion patterns as attractors. At the same time, the robot also records the sensor data. Thus the robot is allowed to have the motion attractors corresponding to the environment. In the teaching playback phase for the autonomous hexapedal locomotion, we use two feedforward neural network sequentially. The first one is for identifying the environment from the senser inputs. The second one is so-colded an associative memory for deciding locomotive motions. Through simulation experiments, we show that the robot is enabled to walk on the flat plane and climb an obstacle autonomously.

This paper deals with the design of CPG for our four-legged robot. Although the robot we built is able to walk, it sometimes fall down, which means that its stability in walking is insufficient. One of the reasons of this problem is lack of a theoretical method for designing the walking pattern. In our study, the trajectory of each toe was manually designed by human, which does not guarantee finding suitable trajectories out. Therefore we focused on the CPG control. In order to apply the CPG control to our robot, a coupling pattern of CPG for a four-legged robot was constructed, which was reported in this paper.

Quadruped walking robots are well known as the robots which have high adaptability to rough-terrain and were able to move while keeping static stability. For those reasons, they have been expected to be utilized on rough-terrain, and then, huge number of studies have also been conducted. However, optimal joint configuration according to the purpose has not clarified yet. Our goal is to clarify the way how to control a quadruped walking robot which has 4 DOF on its each leg. In this paper, we report the results we got by analysis focused on force on each joint of a leg.

This paper reports on the application of Symmetric Walking Control to a torque-controlled hydraulic quadruped robot RL-A1 having 1.1 m height and 75 kg weight. By maintaining controlled symmetry between the supporting legs and swinging legs as viewed from the vertical plane, we can expect the robot does not fall even under strong perturbation. The idea is verified through the exeriments where the robot can stably take a step or continuously walk on the ground even a large perturbation is applied.

This paper proposes a foldable legged robot that is able to move in a narrow environment. The robot produces reaction force on a wall at each leg when the legs are folded by external pressure. The robot is based on a Roll-free all sides legged robot. Since this robot has a gait at every face to move toward the same direction, the foldable robot is able to keep moving to the same direction even it enters into narrow environment. A series of experiments were performed by using an experimental machine, which confirmed the possibility of moving in narrow environments by using legs.

In Japan, infrastructure intensively built after the period of high economic growth rapidly deteriorates in the future. We developed robotics system for careful inspection the integrity of structures and prevent emergencies. This system includes 3 robots and camera that install under the bridge or on the wall. The robots move by rope that installed parallel to the building site and inspect by camera. However, these robots oscillate and rotate by influence of the wind. Therefore, we developed the three dimensional gimbal head that install in the robot. This system necessary for move the camera, that inspecting a wall or a bridge.

In the seismic hazard generated some time ago, the inspection robot was expected of the information gathering to the distresses region where nobody was able to enter. However, because the robot technology to the disaster correspondence is not continuously prepared, the work result by the robot is not achieved enough. In our aim of study, a new robot inspecting the building wall and the estimated method of deterioration that can be used even due to the normal work and the disaster are developed. In Part 3, the element modules and the whole idea of the inspection robot to detect the wall in the building at time usually are reported. The inspection robot is composed of various modules, and this robot can be moved on unleveled land and to climb even by the vertical wall thought reaction force by duct fan and hung wire.

Most of the construction work of the building facilities occupies the high place work using the temporary scaffolding and the work near the ceiling, securing the safety and improving the productivity are the subjects. In addition to monitoring by a camera or the like, the UAV is required to have a function to stably contact the surrounding environment such as a ceiling and a wall to maintain it in order to perform or support specific work. Therefore, in this research, we will prototype a UAV that performs the air volume measurement work at the air conditioning air blowout port installed on the ceiling surface and examine these necessary functions.

When the walls of structures are inspected by a small robot, footholds are not required and preparation to inspect the walls is easy unlike a large robot. For these reasons, inspection costs would be reduced and human safety would be improved. Therefore, in order to realize hammering inspection by a small robot, we have developed a hammering inspection system that can be mounted on any small robots. Based on frequency components obtained by wavelet transform of hammering sound, we have compared two identification methods, using cross correlation method and using machine learning. As a result, we have confirmed that machine learning is suitable with more precision as an identification method to inspect the walls.

In the post-disaster restoration works, many remote and unmanned construction technologies have been used. Recently, disasters at river edge or shallow water place are increasing and remote and unmanned construction technologies which are effective in these environments are expected.
We are joining the project that aims to develop a remotely controlled heavy carrier robot, and the unmanned construction system using the developed carrier for the reduction of disaster damages and for the facilitation of early restoration at shallow water environment.
This paper describes developing of prototype machine for the remotely controlled shallow water heavy carrier robot. The carrier needs to be amphibian for extending its reach. The heavy carrier robot was able to drive in 1.5m depth water by remote operation.

In nature there are many insects with obvious differences in the length of each leg. However, there was not much discussion about the influence of these leg length differences on walking. So in this research, in order to elucidate the influence of the difference in leg length on walking, we focus on locusts, whose hind legs are considerably longer than other legs, and developed a six-legged walking robot based on the structure of locust. The developed walking robot is controlled by changing leg trajectories according to the installation state of each leg. We conducted walking experiments with the developed robot while varying the length of the hind legs and confirmed that leg length affects walking speed of the robot. These results suggest that there is a length of the hind leg suitable for walking.

Recently, legged robots have been developed to work in a dangerous space for an investigation, rescue, and so on. This paper introduces a method for a Limb-type robot to move in frame structures such as a ladder or jungle-gym. The proposed method is based on path planning and motion generation. For path planning problem, A* algorithm plans a path while taking the workspace of limbs of the robot in consideration. The robot can then move along the resultant path, and the robot need to stop to hold new supported position on the way. The velocity change for stoppages is decided by mechanical constrains of the robot. The motion is generated from the velocity by Inverse kinematics. After the method is explained, kinematic simulation is conducted, and some problems and the improvement plans are indicated.

This paper reports motor control for running and walking of "YANBO - II",which is a bipedal robot with reduced degrees of freedom. YANBO - II is composed of eight degrees of freedom, three degrees of freedom for determining position in space, three degrees of freedom for determining posture, and two degrees of freedom to control the center of gravity within the support area on the foot. In addition, YANBO - II is equipped with an infinite rotation motor on the soles of the feet, and running by using the sole as a wheel is possible. Proportional control of the joint motor to make the YANBO-II walk and the implementation of running and walking will be explained.

In this report, the simulation model of the bipedal robot with reduced degrees of freedom and considerations of the sensors mounted on the sole of the robot are explained. A multi-legged walking robot has usually many degrees of freedom and also many control parameters. This research aims at developing the legged robot has less degrees of freedom to achieve the mission in the real world to make its control easier. When the robot walks autonomously, it is quite important to recognize the condition of terrain and contacts with it. It uses sensor to control the attitude and joints of the robot in order to achieve the stable locomotion.

This paper proposes an inexpensive torque sensor using float differential (FD) mechanism for legged robot. Stability indicators like ZMP are provided by floor reaction force. Though most of conventional force sensors for measuring floor reaction force have complicated structure and are expensive. Thus, we designed inexpensive torque sensor using FD mechanism and FSR, cheap pressure sensor. In the proposing design, the sensor can be mounted on outside of RC servo case. In this paper, the design and fabrication of the sensor mechanism and evaluation of its sensing of torque and force vector are shown.