This paper presents cooperative multi-robot map merging process for mobile robots operating in indoor environments. Recent methods in single robot SLAM has been very successful in mapping small to medium sized areas but suffers from optimizing maps for larger areas due to drift. Multi-robot systems can simplify the process of mapping and exploring large environments by distributing tasks between the robot teams. By merging the local maps from the robots, a global representation of the area can be used for different applications. We present an online multi-robot mapping based on multiple pose graphs. The main contributions of this paper is a an online map merging process that formulates relationship between multiple pose graphs using visual and scan features to give a consistent global multi-robot mapping solution. The pose graph information are shared to all robots and a central workstation is used where the local pose graphs are optimized globally. The process allows our system for data association and efficient loop closure even for large areas. A task distribution method is used for complete autonomous exploration in unknown environments. The proposed system is tested with real robots and the results are discussed.

The authors have developed cooperative system of three manipulators that are position controlled. The problem with this system is that excessive internal forces occur due to position error. Therefore, we have used six wires to prevent excessive internal forces. Also, the system using six wires can hold the object rigidly against external forces of a certain size. However, when considering the use of this system for assembly and disassembly work, compliance is required against external forces. Thus, this paper discusses cooperative system using three wires and provides calculation method for positioning of the object.

In previous research, we proposed a new hydraulic system named ”Multi-channel EHA” that can control multi-DOF by controlling a single pump and electromagnetic switching valves, and confirmed its potential experimentally. In the case of multi-DOF driving, the conventional hydraulic systems have a problem of heavy mass and low efficiency. On the other hand, the advantages of this proposed method are lightweight and high energy efficiency. Regarding the system mass, it is much lighter than the conventional hydraulic systems for multi-DOF drive with multiple pumps because the proposed system can realize multi-DOF drive with a single pump. In this paper, we experimentally verified energy efficiency and quantitatively compared it to conventional hydraulic systems.

Robots have a problem of locomotion in extremely rough terrain. Legged robots have a high locomotive ability because it is possible to select a landing point discretely. However, a combination of electric motors and speed reducers adopted in many legged robots has a problem such as low shock resistance. Because of low shock resistance, it is difficult for legged robots to operate motions with high shock such as jump off. Therefore, a leg mechanism which can be applied to legged robots and has high locomotive ability is required. In our previous researches, we developed a hydraulic cylinder for robots. In this paper, we focused on the controllability of this hydraulic cylinder and compared the controllability with the conventional product through measurements of damping performance. We also conducted an experiment for measuring the damping performance with a prototype of the leg mechanism using the hydraulic cylinder.

Edible materials have been utilized for soft robots and made them biodegradable and biocompatible. So far, the authors have devised and realized not only soft edible robotic elements but also hard edible materials essential to tasks such as insertion into a narrow environment. However, hard materials tend to require casting mold and thus have infeasible shapes. In this research, In this research, we have devised and devised an edible flexible rigid switching mechanism capable of switching the rigidity state, which was flexible and stiff, respectively, in the research and development of the edible robot we are working on, Characteristics were confirmed by actual experiments.

This study describes design of low pulsation type water hydraulic motor practicable under a radiation environment. Basic structure of the motor is radial piston type. Reciprocation of 4 pistons was changed to rotary motion by a cam mechanism. The cam is always interposed between 2 cam followers without a spring, and the opposite 2 pistons are connected. The cam shape was designed so that pulsation of the rotational speed and the torque might be zero with the constant flow rate of water supplied to the motor. A distribution valve to change water flow to each cylinder was also newly designed with consideration of low viscosity of water.

This paper proposes a new design method of the helical actuator which can realize a high ratio contraction when pressurized. Based on the same working principle of the old design, a new fabrication way using only the one-way stretchable cloth is proposed to satisfy the situation where tiny size of the helical actuator is needed. A mathematical model is analyzed and set up based on the geometric structure of the new structure of the actuator. Experiments are conducted to examine the validity of the model.

This paper presents motion generation technique for a flexible pneumatic robot based on position estimation of linear actuators. The authors previously proposed an in-pipe robot which drives in narrow pipes with flexible pneumatic linear actuators. In previous paper, a static model to estimate the position of actuator’s sliding parts by applied pressure and flow rate was proposed and basics were discussed. In this paper, an effect of inflation diameter change was introduced into the model to increase estimation accuracy. The model was then integrated into a robot’s motion generation logic, and automated motion generation was tested in horizontal straight pipes.

The pneumatic soft actuator converts the input energy by utilizing structural deformation or material deformation according to the applied pressure to obtain the output. Although various forms of operation are conceivable, in practice it was difficult to freely design under manufacturing constraints. In recent years, with the development of 3D printers, flexible materials such as rubber as well as hard materials such as conventional resin can be printed. It is now possible to manufacture an actuator having a complicated shape combining a resin material and a rubber material that have been difficult to produce. In this paper, Pleated Pneumatic Artificial Muscles using 3D printer which can design freely using resin and rubber material with fabricated and evaluated contraction characteristics.

The conventional mechanical structures are composed of many kinds of characteristic parts; those are complexity. The authors propose to manufacture the machine with uniformed structure of simple parts. We fabricated novel soft mechanisms that are composed of only woven strings and thin McKibben muscles. The structure is uniformed; nevertheless, the forms and movements of mechanisms are diverse. In this paper, we report the variety of deformation of active textiles that are woven in three foundation weave with thin McKibben muscles and some strings.

Electro-Hydrostatic Actuator (EHA) has high backdrivability and high control bandwidth, but it doesn’t have sufficient power as a hydraulic actuator. For the purpose to be compatible disturbance flexibility and high power, we developed a high power robot arm named“ Hydracer ”driven by EHA. For the robot arm, we developed a ceramics enhanced involute gear pump, linear EHA module composed of the pump and 3D printed cylinder and ceramics enhanced single vane motor to enhance power and pressure of it. The main 3 joints of the robot arm apply the oscillating slider crank mechanism driven by 2 cylinders. The wrist joint of the robot arm is driven by 3 ceramics single vane motors in series. The flow paths of the wrist joint are manifolded and the wrist joint became compact by arranging three pumps and one swivel joint.

This paper describes the development of a new valve system that switches the pressurized modes of three or more chambers by self-excited vibration instead of using electricity. Recently pneumatic soft robots developed actively which can move narrow environments by taking advantage of its flexibility. Generally, solenoid valves are required to drive pneumatic soft robots, but they are losing the original flexibility of the soft robots and also makes robots difficult to be used in the environment where explosion-proof performance is required. Therefore, we developed a new valve and checked its function by experiments. Finally, the validity of our proposing method was verified by implementing it on a pneumatic mobile robot generating traveling wave and realized the propulsion by single supply line without electricity.

When a joint of human body doesn’t move for a long time, an ability of the muscles around joints will be lost and joint contracture occurs. This rehabilitation device has two sticks that rotate and translate by double lack and pinion mechanism, which is driven with pneumatic cylinders. The sticks can stretch finger’s joints and wrist, but, the shape of finger and hand can not be recognized. In this study, we detect the position of finger’s joints and wrist joint by USB camera and OpenCV graphic library and design effective rehabilitation by the position. Methods of detection are detecting marker of joints and corner of approched hand’s outer shape.

High power, light weight and flexible powered suits are required. However, they are not actually made. Because conventional hydraulic actuators are high power, but they are heavy and hard because they are made of metal. So it is not suitable for powered suits aiming at light weight and flexibility. The McKiben artificial muscle is composed of a rubber tube on the inside and a sleeve knitted with chemical fiber on the outside, so it is very lightweight and has a flexible structure. It is suitable for powered suits. We developed thin Hydraulic artificial muscle (THAM) and measured its static characteristics to demonstrate the shrinking force of 1262 N when applying 5 MPa. We also fabricated an active textile woven using THAM to investigate the applicability to powered suit.

Long flexible continuum robots have a high potential for search and rescue operations that explore deep layered debris. Air Floating Active Scope Camera (ASC) have been developed for search and rescue in a narrow space. However, it is confirmed the phenomenon that sticks to the ground when floating. Air jet nozzle with air curtain have been developed to solve problems. In this paper, we measure forces occurring in the direction of the ground, and determine parameters for air curtain nozzle from the measured force.

We propose a “Polygonal Origami Actuator”: rigid foldable bellows mechanisms that consist of hinges and plates and contract by fluid power. Several pneumatic artificial muscles have been proposed for a linear unidirectional motor, however, the power performance has been limited due to their dead space and the soft material. In this paper, two bellows mechanism which contracts by pressurizing the inner chambers enables high power and compactification are shown. The kinetics of the proposed two bellows actuator was studied from the geometrical point of view, and the basic characteristics of the bellows such as volume and the contraction ratio are analyzed.

In Japan, 27.7 percent of the population are elderly people and these trend of aging will be advancing. Along with the aging society, the number of people who are difficult to walk by oneself is increasing. Walking is the important activity in their daily living. Without walking, not only the athletic ability decrease but also they are possible to develop a lifestyle-related disease. Therefore, the demand for a walking rehabilitation device is increasing. Walking rehabilitation supports the lower extremities such as ankle, knee and waist but this reseach focuses on the ankle. Introducing pump and PC control, high functional rehabilitation is expected.

Some diabetic patients complain diabetic paw lesions, which may result in ulceration or gangrene on the foot or in the worst case forced to cut the foot. One of the causes of foot ulcer formation is plantar pressure abnormality due to limitation of range of joint movement. In this study, in order to prevent foot ulcer formation due to diabetic foot lesions, walking support shoes that actively support dorsiflexion action of the ankle joint are produced and the effect is verified. Generally, with such active support, it is necessary to carry an energy source separately, which leads to large-sized and heavy equipment. Considering the use in everyday life, instead of using another power source, we support the dorsiflexion by the weight of the wearer. In addition, for a diabetic patient concurrent with neurological disorder of the lower limbs, a device for biofeedback of the sensation of the sole is manufactured and the effect thereof is verified.

This paper proposes an experimental result about real-time recognition of the position and the orientation of traffic lights based on this Photo-model-based method in a simulated environment. Because experimenting at real traffic intersections can cause a lot of trouble for pedestrians and vehicles, we have built a road simulation environment. The Photo-model-based method can obtain the current position and posture of the object by using only two calibrated cameras through the image of the object that has been acquired. I will introduce the fitness function of this method and the method of exploration using genetic algorithms in this article, as well as the latest data in this environment. This time, the results of the experiment were obtained by experimenting with traffic lights in red states. The experimental results in real time will be reported.

Real-time kinematic (RTK) positioning is a GNSS positioning technique that can provide precise position within several centimeters of accuracy. In general, it is more difficult to solve GNSS carrier-phase ambiguities using RTK-GNSS in multipath environments with a low-cost single-frequency GNSS receiver than with an expensive dual-frequency GNSS receiver. The purpose of this work is to improve the accuracy of single-frequency RTK-GNSS positioning. The proposed technique consists of two approaches. The first approach is the autonomous selection of the GNSS signals used for RTK-GNSS positioning. We select satellite signals depending on their quality, which can be identified by signal strength. The second approach combines the GNSS Doppler velocity to resolve the integer ambiguities. To estimate more precise initial position and covariance of RTK-GNSS, we use the previous fixed solution and the Doppler velocity. Extending the open-source RTKLIB software for RTK-GNSS positioning, we confirm the effectiveness of the proposed technique.

Automated driving reduces the burden on the driver, however also makes it difficult for the driver to understand the current situation and predict the future movement of the vehicle. To facilitate the prediction of the future behavior of the vehicle by the driver, this paper aims to design and evaluate a haptic interface that actuates the vehicle seat. Our system displays to the driver the movement of the vehicle a few seconds in the future, which allows the driver to make predictions and preparations. Using a driving simulator, we compared the conditions where the movement of the car was displayed in advance for the length of different time. The subjective evaluation of the driver showed that the predictability of the behavior of the vehicle were significantly increased compared to the case without display. The experiment also showed that comfortable feeling significantly decreased if the preceding display is too early.

Self-driving car needs accurate position, and a highly accurate map is indispensable. However, it is expensive to create and maintain. Therefore, we proposed an image updating method using matching between ortho images generated by a rear view camera and aerial ortho images. Aerial ortho images have already been created in a wide range, and we believe that the proposed method is useful for partial updating of such images. In this study, from the experimental results, we confirmed that it is possible to update and generate high resolution aerial ortho images without obstacles such as vehicles and traffic lights by using rear view camera.

In this paper, we describe accuracy evaluation of 3D point clouds applied to autonomous driving. An error due to measurement may remain in the constructed 3D point cloud. Also, in the case of using for localization, the problem that localization failed is known due to the place where the continuous structure. The purpose of this study is to examine the evaluation index of the map and evaluate the accuracy of the 3D point cloud. In the proposed method, we evaluate the map from the viewpoint of Localization and propose an evaluation method using convergence by matching. By the proposed evaluation method, it was found that the place where Localization is possible and the place where the error remained in map are clearly separable. Moreover, we think that the continuous structure place can be classified from the tendency of the convergence point.

When highly-automated vehicles operating in level 3 encounter system limitations, they issue a takeover request (TOR) to transfer the control authority from the automated driving (AD) system to a human driver. In such ‘unscheduled’ situations, the driver needs immediately re-engage in the driving task both physically and cognitively. In this study, we developed a driver situational awareness estimation system based on glance information. We collected a lot of glance data when both safe and dangerous takeover situations by using a driving simulator, and we derived the standard glance model including the glance area and time. To evaluate the effectiveness of the model, we created a glance guidance system to visually indicate regions with insufficient glance. We found that the guidance system improved driving performance and reduced the number of accidents.

To make full use of the advantages of automatic driving while retaining the enjoyment and freedom of the manual driving, we have proposed a tactical-level input (TLI) that allows the driver to input dynamic driving tasks such as parking and turning. In the future state input using touch interface, eye movement between the interface and front is troublesome and it takes time to operate the screen. In the command input using haptic interface, it cannot reserve input to a distant place and the amount of information is small. To solve these problems, we develop a visual-haptic interface by integrating them, which has a planar-movable haptic device and an input/output state indicator using transparent screen. The results of experiments using a driving simulator indicate that the proposed interface solve the above disadvantages while remaining the advantages.

Technological innovation of AI, IoT and so on in recent years is called the Fourth Industrial Revolution, and the development of autonomous driving is thriving. However, in the rapid progress of technological innovation, if cars in the world are automated, further lack of IT engineer is expected. In this research, the authors develop a small autonomous driving hardware platform for education and research. It mainly consists of ordinary consumer electronics such as Raspberry Pi and also is an open source project. Thanks to these two features, this platform is not only affordable but also easy-to-modify by everyone.

In this paper, we will also consider methods to estimate of the position with high accuracy even in urban areas. We use satellite positioning by GNSS for position estimation. RTK with high precision in satellite positioning can estimate position with accuracy of cm class. However, in urban areas, multipath causes degradation of positioning results. Therefore, we propose a method to improve positioning results by improving robustness against multipath by utilizing vehicle trajectory. In the proposed method, we improve the initial conditions when estimating the float solution in RTK. Conventionally, this initial condition suffered from multipath and a large error occurred. However, in the proposed method, since the robustness to multipath is improved, the error of the initial condition is small. Accuracy of RTK positioning improved by using the initial condition that suppresses multipath.

There are many path planning methods using geometric curves as a reference path. It is important to construct smooth reference path because the efficiency and ride comfort of the vehicle depend on the smoothness of the reference path. Moreover, in the case where the vehicle speed is changed, we should consider the fact that the faster the vehicle speed is, the more effort is required to follow the reference path. Hence we use the kinematic model of vehicle to construct the path for considering speed change of vehicle. We propose the method to generate the optimal path minimizing the square sum of the vehicle lateral acceleration. The optimal paths can be easily obtained because the optimal control problem is reduced by the change of the independent variable and normalization. Additionally, the optimal paths can be summarized in some figures or so called field of extremals.

In mountainous region and underpopulated area, the lack of means of transportation for impaired people are considered as a social problem. In order to solve this problem, Center of Innovation project (COI) conducts research on low-speed autonomous vehicle as a mobility service. However, low-speed autonomous vehicle tends to disturb traffic because of maximum acceleration / deceleration and cruising speed limitation. The objective of this study is to construct speed planning and control system of low-speed autonomous vehicle at intersection considering surroundings. The proposed system searches optimal acceleration using potential field based on estimated time to center of intersection. The effectiveness of the system is verified by simulations.

In conventional vehicle, the noise generated by engine mainly deteriorates cabin comfortability. However, in ultra-compact electric vehicle which does not include engine, an interior noise caused by road and wind noise have a large impact on user comfort in the interior space. In previous study, we propose a system whereby giant magnetostrictive actuator is installed on the roof of an ultra-compact electric vehicle to actively cut interface vibration noise, i.e., road noise, transmitted to the interior of the vehicle and we aim to reduce the noise level in an interior of a vehicle using this proposed active noise control system. Furthermore, we studied fundamental consideration of masking. In this study, we focused on 1/f fluctuation for masking music and evaluate the vehicle interior comfort by using the brain waves of the crew, in the case of operating driving simulator.

Self-driving systems can be classified into those based on object recognition and motion control methods and those based on End-to-End learning. This research aims to develop a learning model that generates vehicle control signals only from camera images in an End-to-End method. In order to collect training data to be used for deep learning, we constructed a simulation environment where we can drive a vehicle using CARLA driving simulator. We collected as a dataset pairs of images from a camera mounted on the vehicle and human operations. We use Convolutional Neural Networks and Recurrent Neural Networks for learning and compare the models.

In recent year, high efficiency of automobile engines has been demanded for one of energy saving. However, the surging of valve spring and valve bounce are occurred by mechanical valve system. Therefore, the innovative technologies of efficiency of engine has to be established urgently. We design the system of engine valve which is capable of opening and closing by using linear motor. Our proposal valve system is able to be combustion efficiency improvement and decreasing pumping loss by arbitrary valve lift and timing. In this study, we focus on magnetization of permanent magnet installed in linear motor and consider electromagnetic characteristics to create a new valve system.

This paper presents a driving behavior modeling method using autoregressive input-output hidden Markov models (AIOHMMs). First, model parameter learning and discrimination ways using the AIOHMMs are detailed. In experiments, we model four driving maneuvers with driver’s eye-gaze and ego-vehicle localization information and compare maneuver discrimination performances by four types of HMMs. Additionally, we discuss modeling performance by the AIOHMMs through the experiments.

In this paper, as the target to predict reverse run around a highway merge, in order to judge whether the driver behavior is an action leading to reverse run, it is described the method to express the driving situation in binary coded strings based on various sensor information. In the two situations assumed in this study, a reverse run behavior template is created and template matching is performed between the encoded strings and its template. The encoded string is generated from the presence / absence of vehicles around the subject vehicle, road sign and acceleration / deceleration information of the subject vehicle. Verification experiments reproducing the vicinity of the highway junction within the school will be conducted to examine the accuracy and effectiveness of the generated coded string.

In this report, a drive-log recorder for bicycle is introduced. This is similar to drive recorder for vehicles. There are few device that can alert to cyclists based on running state in real time. The recorder has two acceleration sensors, rotary encoder, and camera. The effectiveness was confirmed by measuring signal of each sensor installed to the bicycle.

An automobile is equipped with many Electronic Control Units (ECUs) to which various sensors are connected, and advanced control is achieved by exchanging information among the ECUs through automotive networks. However, it is difficult for beginners just like students to implement such networks on short development schedule, because costs for designing, constructing, testing and managing the networks are a burden to them with an increase of connected nodes. For this problem, we have proposed a framework of plug and play connection in which sensor nodes are automatically connected to automotive network. In this paper, we propose a multichannel switcher, a sensor I/F judgement algorithm and a GUI environment for setting sensor-specific information to put the framework to practical use.

Trajectory planning is an important technology to make autonomous driving comfortable and safe. This paper is focused on trajectory planning in the situation which there is a leading vehicle to follow. Under such circumstances, autonomous vehicle must maintain a safe distance to the leading vehicle. However, a trajectory considered a safe inter-vehicle distance has a risk to brake excessively when a cut-in vehicle is detected. Excessive braking gives a bad influence to traffic flow behind and comfortableness for people on board. Therefore, this paper proposes the method to generate a trajectory to keep an optimal distance to the leading vehicle and control deceleration in the situation to be cut in by other vehicle.

In this paper, we propose high-speed tracking system using machine learning and image processing to identify overlapping vehicles. Since existing tracking system can track only one car and fail to track overlapping cars, it is necessary to be able to identify overlapping cars and continue tracking. We adopt deep learning algorithm YOLO[1] to identify overlapping cars and high-speed tracking system using correlation filter and IPM method for tracking. Our proposed method implemented high-speed and high-precision tracking system for overlapping cars.

For personal care robots that coexist with people in public or living spaces, it is necessary for robots to be safe as machines, such as following ISO 13482 / JIS B 8445 standards. However, when using a robot, it is insufficient for robots to be safe only as a machine, and it is important how users operate and manage such robots safely. Volunteers have discussed this at RRI (Robotics Revolution Initiative Council) and after another year discussion at the JIS standard committee at JARA, a new standard is to be issued, so we explain and discuss about the standard.

In the security services, safe improvement using the robotic automation is demanded from lack of workforce and the increase in tourist by the Games of the XXXII Olympiad. In this study, we developed the robot system using the autonomous mobile security robot and the fixed point cameras. In this paper, we show a process of the risk assessment that assumed a railroad station and a part of the risk assessment sheet, and the development of the security robot "Perseusbot". Finally we introduce the details of the demonstration experiment in the railroad station.

Over the past years, Japan has become a low birthrate country with an aging population. Hence, the demand for life support by robots and robotic technology has increased. We aim to create a society that is seamlessly connected from home to town by cooperating with senior cars, forming a mobility and robot network. As a new form of collaboration between robots and mobility, we developed a system in which the robot follows a senior car being automatically driven. By having a robot follows a senior car, the robot is expected to carry baggage that cannot be held, watch over the elderly driver, and so on. We attached an augmented reality (AR) marker to the senior car, which the robot recognized with a web camera, allowing the robot to follow the senior car successfully. The experiment demonstrates that the robot can follow the senior car at a distance of 1 m.

Recently, the development of human-friendly robots has been desired. Wheeled mobile robots are often used in human friendly robots. When developing the human-friendly wheeled mobile robots, safety consideration for humans is one of most important issues. Computer-aided control techniques can improve the safety of the robots. However, if the robot’s computer breaks down, the robot will be dangerous for humans because the robot can move unintentionally. Furthermore, the robot may move down a slope at high speeds after the batteries in the robot have died. To solve these problems, we have developed a velocity-based mechanical safety brake. In this paper, we describe the design of a wheeled mobile robot equipped with the velocity-based mechanical brakes to verify the effectiveness of the brakes by experiments.

In this research, we describe the design of a velocity-based mechanical safety brake for the wheeled mobile robots. If this safety brake detects an unexpected drive-wheel’s angular velocity, it switches off all of the robot’s motors and then applies brakes to the drive-wheels. The safety brake works even when the robot’s computer breaks down, because it consists of only passive mechanical components without actuators, controllers, or batteries. The safety brake has two rollers and two mechanical velocity-based lock devices. Firstly, we introduce the features of the mechanical safety brake. Secondly, we explain the structure of the mechanical safety brake. Finally, we describe the structure and mechanism of the mechanical velocity-based lock device used in the mechanical safety brake.

The passive dynamic control (PDC) is a mechanical system control method based on an inherently safe design means using a brake mechanism effectively. Since the PDC follow-up control can be envisaged as gradually narrowing down the wide area around a control object to enclose it toward the target trajectory, the authors designate this control as “enclosing control.” Here, an experimental setup with an MR brake attached to a rotary pneumatic actuator is equipped, and the effectiveness of the enclosing control is verified by the experimental setup. The MR brake is a braking device with high braking torque and excellent responsiveness.

We have developed the human friendly robot that consists of two arms, a trunk, a waist, an omnidirectional mobile base. Each arm of the robot has 2 DOF shoulder, 1 DOF elbow and 2 DOF hand, and its waist has 3 DOF. Its height is 1.42[m], its arm length is 0.76[m], and its total weight is 83.5 [kg]. In this paper, a new passive collision force suppression mechanism is described, which consists of two release air bags, a clutch gear, two coil springs, and four liner guides. Also, a new soft cover is developed to cover the entire outside of the robot. When the collision occurs between the robot and an object, the joints of the arm and the waist are disconnected by the collision force suppression mechanism depending on the magnitude of collision forces. Collision experiments are conducted, and the effectiveness of the collision force suppression mechanism is verified.

In the field of robot-human collaboration, while safety verification data are already provided in ISO/TS 15066 as informative data, a testing method for validation is not established. In our study, we design the dummy hand the mechanical property of which is close to that of human to propose the testing method using the dummy. In order to design such dummy, we estimate human physical parameters by conducting dynamic experiment in which three points in each hand of the subjects were clamped as demonstrating the worst case of impacting a hand as the first step. For conducting dynamic experiment, we design impact force measurement device in consideration of safety enough. As a result, we measure transition of impact force, and estimate human physical parameters under various conditions.

We develop the Autonomous guide robot designed on the basis of intrinsic safety. As a serious risk of a mobile robot, we have to consider runover. In order to reduce the severity of harm of runover, we have to reduce the weight of the robot. But we have no design reference value. In this paper, when the robot overruns against the foot of young child, in order to derive the load value where the fracture trouble occurs, we estimate it to use the past accident data and report on the derivation process.

This paper reports a development of impact test for UAV (Unmanned Aerial Vehicle) in order to understand injury level caused by free-fall UAV. Test results given by originally created head impactor devices are shown. The ordinal criterion for head injury: HIC value were measured.

The main objective of our research is to construct an evaluation scheme for a use case of mind drone interface (MDI). Moreover, in order to judge whether the evaluation scheme is appropriate, we would like to consider the safety margin of MDI at the time of approaching the crisis. Several research work and projects have already proposed MDI, but still it is the work of the laboratory environment with safe case. To install MDI practically, we develop the evaluation scheme about the difference of reaction time on the MDI between safe or crisis condition. In this research, we construct an example system of MDI with extracting reaction duration of mind to move actual drone. We also evaluate the difference of reaction duration between safe or crisis condition in the laboratory environment. From this result, we can notice enough distance from drone to obstacles is need to add 7 seconds to control-loss time.

One of the serious hazards is a fall to the wearer of the wearable mobile support equipment. Preliminary gait analysis will be conducted with the aim of grasping the sign of falling at an early stage in order to prevent falls. Various walking patterns were performed on subjects who simulated the elderly using two types of walking assist equipment.