Accurate estimation of video pulse wave is difficult due to the influence of ambient light and body motion. In this study, with the aim of improving the measurement accuracy of heart rate and other parameters by analyzing heart rate variability using video pulse wave, we compared data measured under multiple lighting conditions and pulse wave extraction methods with data obtained from ECG.

In the decommissioning of Fukushima Daiichi Nuclear Power Station, the modularized rail structure, “ Rail DRAGON ” has been developed to observe inside the PCV for avoiding the troubles of robot teleoperation. Because there are obstacles on the path of the extending rail, the operators need to control the joint angles and avoid them. In this research, we propose “ wire switch method ” to concurrently execute the module connection and the control of joint angles. By setting up a detour of the power line and automatically judging the suitable moment that the fluctuation of angle signals is small, we proved that joint control can be maintained during the wire switching.

This paper outlines the specifications for a heavy-duty robot designed for decommissioning of the Fukushima Daiichi Nuclear Power Plant and proposes a mechanism configuration addressing the need for a broad range of arm motion, the ability to lift heavy objects, ease of decontamination, and high radiation resistance. The proposed configurations includes a 3-DOF pantograph arm with a pneumatic balancer for heavy object manipulation, a 4-DOF wrist with a remote-driven mechanism for enhanced radiation resistance, and a 6-wheeled mobile mechanism with adjustable tire pressure for ground pressure control. A robot with the proposed mechanism configuration, Phoenix-H, was fabricated and examined. It was confirmed that Phoenix-H can handle a 150 kg weight and achieve improved turning speed through ground pressure distribution control via tire pressure adjustments.

This paper proposes a simple and practical mechanism for winding cables without the use of slip rings. In this mechanism, the cable fed from the cable drive is guided by a rotating guide and placed in a curled-up shape. This disperses torsion and eliminates slip rings. In addition, the guide is propelled by the reaction force from the cable to achieve a simple mechanism with a small number of actuators. This paper also describes the constraints on guide geometry and the guide geometry design method based on the constraints. The prototype machine was confirmed to be capable of winding long cables in an orderly manner and discharging them without twisting.

This research proposes a long-reach robot arm with high aspect ratio, high retractility and multiple degrees of freedom for investigations of an inside of a reactor pressure vessel in Fukushima Daiichi Nuclear Power Plant. In this paper, we propose a new mechanism for the long-reach robot arm that extends vertically while bending in the middle by driving wires outside of a telescopic rod that is extended pneumatically, which is extended the Truss Arm mechanism proposed by our research group. In addition, the feasibility of the proposed mechanism is verified through experiments using a scale model prototype with the same configuration as the full-scale model.

In this study, we proposed an intuitive viewpoint change method that maintains consistency between the visual system and the operation system in a remote control system with multiple viewpoints including a free viewpoint, and verified its effectiveness on an experimental system that simulates a decommissioned nuclear reactor environment. In the previous study, the viewpoints were fixed except for the hand camera, but in this study, a free viewpoint camera was newly introduced to allow the operator to freely change viewpoints during the operation.

In this study, we are developing a winding type high rigid long arm using low melting point alloy for the investigation of the inside of nuclear reactors. In this paper, we describes the principle of a temperature measurement device using changes in the resonant frequency of LC resonant circuits. Furthermore, we design a prototype of an angle changing device, and confirmed the performance of it.

We are developing a system to reflect and visualize the measurement results obtained from small directional radiation detectors on a 3D point cloud environmental map. This paper describes a method to identify the location of radiation source distribution from the measurement data collected by multiple robots equipped with the detectors and to reflect them on a map.

This paper describes the development of a remotely operated robot for decontamination of high places in a nuclear reactor building, which participated in the 8th Creative Robot Contest for Decommissioning. The robot was designed to perform a complicated course transportation and decontamination of a contaminated area at a height of 2.7 m by remote control. Therefore, the omni-directional moving mechanism of the mecanum wheel provides high mobility, and the pantograph mechanism provides high deployment capability in spite of its small size. In addition, the telescopic arm mechanism enables precise decontamination work over a wide area. Moreover, a user interface has been developed for safe operation of the robot.

This paper describes the development of the user-interface for intuitively maneuvering remotely operated robots. In the remote operation, the communication delay makes difficult to maneuver the robots, and it will affect operator’s mental workload. Reducing the mental workload is important factor for safe task execution. To reduce the operator’s mental workload, we focused on improving the operator’s Sense of Agency.

This paper describes an on-line 3D environmental modeling system from time-series images for monitoring the situation inside Primary Containment Vessel of nuclear power plant. The prototype system has been designed and implemented to sequentially execute the processes of remote image capture, image transfer over network and 3D modeling calculations. The demonstration test combining with a long-reach bendable modularized rail structure and monitoring robot was conducted. We report the experimental result and discuss future challenges.

This paper describes the evaluation of applying several feature keypoint extraction methods, including deep learning-based method, to video images containing effect of imaging noise due to radiation, turbidity of water, and suspended matter acquired by the remotely operated robot in nuclear reactor of the Fukushima Daiichi Nuclear Power Station. The results show the feature keypoint extraction method based on deep learning is capable of performing high-density and wide-area 3D reconstruction outperforms conventional methods.

This paper introduces an image quality assessment on a 3D modeling to grasp the internal state of the nuclear reactor for the decommissioning of the Fukushima Daiichi Nuclear Power Station. We tried a quantitative evaluation of videos or images acquired by the investigation inside the primary containment vessels for an efficient 3D reconstruction and investigated the effects of image correction on feature extraction and matching using the images added noise. As a result, we confirmed that the number of corresponding points by feature matching between images improves with an increase in a score of image quality assessment.

This research aims to develop a disaster victim search system without the Internet connection for rescue teams by using infrared cameras for person detection, GPS for precise location tracking, optical flow sensors for accurate positioning, and LiDAR distance sensors for obstacle avoidance. Experiments were conducted using the proposed system to search for victims, and the effectiveness of the system was confirmed.

There is a type of flying robot called a flapping flying robot. By controlling the flapping angle, the center position of flapping motion, and flapping frequency, servo-type flapping flying robots achieve not only flapping flight but also various flight forms such as ascending/descending, turning, flipping, and gliding. In this research, an IMU sensor is equipped to a servo-type flapping flying robot, and we try to achieve hovering flight by changing the pitch angle of the flapping robot by controlling the center position of flapping motion.

For the development of drones, which are a good example of Cyber-Physical Systems where the information world and the physical world are integrated, it is essential to introduce a simulator that can easily verify the consistency of the plant model. In this study, we propose a drone simulator for PX4 that utilizes the open-source Hakoniwa technology. The strengths of the proposed method include flexibility of the plant model and visualization engine, the ease of expanding the functionality of the sensor model, the automatic testing system, and parameterization of the external environment and drone characteristics. We implemented the proposed method and demonstrated that the simulator can realize textbook physical equations and simulate its behavior as a verification of its operation.

Quadrotors equipped with tiltable rotors are overactuated, enabling them to maintain any poses in the air. However, nonlinearity in body dynamics and delays in tilting servos arise when transitioning between poses, presenting challenges to traditional PID-based controllers. In this work, we propose a novel control approach for tilting-quadrotors based on nonlinear model predictive control (NMPC), which leverages the full nonlinear model as the control model. In addition, the servo delay is explicitly considered inside the NMPC. To evaluate the performance, the proposed method has been implemented in Gazebo simulation to track pose setpoints, demonstrating smoother and more rapid pose control.

Modular self-reconfigurable robots (MSRRs) offer enhanced task flexibility by constructing various structures suitable for each task. However, conventional terrestrial MSRRs equipped with wheels face critical challenges, including limitations in the size of constructible structures and system robustness due to elevated wrench loads applied to each module. In this work, we introduce a Aerial MSRR (A-MSRR) system named BEATL, capable of merging and separating in-flight. BEATLE can merge without applying wrench loads to adjacent modules, thereby expanding the scalability and robustness of conventional terrestrial MSRRs. This paper provides an overview of the BEATLE system configuration and, in particular, details the multi-aircraft coupled flight control framework based on contact force control. We also introduce the actual implementation of the prototype and validate the robustness and scalability of the proposed system design through experiments.

This research aims to develop a Detect and Avoid (DAA) system that utilizes omnidirectional cameras as sensors. This study focuses on quantifying scenarios in which UAS must take evasive action based on images captured by omnidirectional cameras. Initially, other UAS are detected from images captured by the omnidirectional cameras. The detected UAS are then tracked through the omnidirectional cameras, and behavioral data of the UAS within the images are collected. Based on this data, we explore the possibility of determining the presence or absence of collision risks. The results show that it is possible to determine the presence or absence of collision risk from the data detected by the omnidirectional camera. However, it was necessary to set the detected distance and the speed of the UAS in order to indicate evasive maneuvers. In addition, the UAS and sensor would have to be quite close to detect prominent features.

As social infrastructure rapidly deteriorated during the period of high economic growth, efforts are being made to improve the efficiency of inspections. Therefore, in recent years, inspections using multi-copters have started to be carried out, but general-purpose multi-copters have some problems when inspecting in narrow spaces or near walls. Therefore, in this paper, we propose a multicopter that can perform multifaceted inspections of infrastructure, and propose a multicopter that incorporates a telescoping and rotating arm mechanism. We developed a simulation model of the proposed multicopter and impremented it.

This paper discuss the image of robot and human in the future and reports the demonstration conducted as a proposal for the vision of robots that coexist with an user and encourage positive behavior. As the progress of aging society, we propose that the robot coexist with an user at home is required to have functions: estimating an user mental or physical state, and advising the user to do appropriate behavior so that keep/improve their mental or physical state. We conducted demonstrations to recreate several scenes as a proposal for the future coexistence of robots and humans at home. This paper also reports about the constructed robot system and the results of generated robot behaviors.

Playful Learning, defined as learning through play, involves two different types of play; free play and guided play. These two types of play during childhood are important for the developments of a healthy mind and body. Playful learning should be provided adapting to characteristics of a child, and toys with interactive functions are used for effective playful learning. We then propose an interactive ball-like robot as a novel interactive toy, which effectively provides free play and guided play adapting to the characteristics of children. This robot is designed to demonstrate three different motions; continuous jumping, left-right turning, and translational movement. These motions are generated by two eccentric motors. In this paper, we present control system for the interactive ball-like robot, which provides guided play to a child, and child-robot interaction experiments using it. The results show that the interactive ball-like robot can induce plays and other physical activities in the children.

As the increase in the number of elderly people in Japan, we have to develop various types of smart device applications for self-care to health promotion. Basically, it is difficult for elderly people to continue daily exercises if the elderly people don’t know the effect of daily exercises. Therefore, we should develop the evaluation method of daily exercise. In this paper, we propose a clustering method based on growing neural gas to evaluate gymnastic exercises by smart devices.

In this research, we develop an owl-shaped pet robot that can communicate with humans. This robot has two degrees of freedom in its neck and has a blinking mechanism. It also recognizes humans and the environment using its built-in pressure sensors and camera and communicates with humans through motions by mechanisms and birdsong by speakers. Furthermore, by autonomously generating emotions even when there is no human interaction, it is possible to create an impression that is more like a living thing. We verify the effectiveness of the developed robot by verifying the effectiveness of emotion presentation and changes in emotion while a long time.

In recent years, the global food crisis has intensified, posing a challenge to secure sustainable food supplies. Factors such as population growth, climate change, and labor shortages in primary industries threaten the stability of food production. In response, insect­based food has gained prominence as an efficient and sustainable food source, offering benefits such as resource efficiency, environmental friendliness, and high nutritional value. This study explores the capability of a large­scale language model (LLM), like ChatGPT, to learn safe cooking methods for insect­based food. Despite ChatGPT's general ethical constraints, we utilize the Fine­tuning API from OpenAI to train the model using a dialog­based dataset. Information about insects and poisons is collected from Wikipedia for training purposes. Evaluation of the model's output is performed using OpenAI's evals, specifically examining the inclusion of appropriate cooking methods. The study assesses the model's performance over 3, 10, and 30 epochs for 50 different types of insects.

We developed a module for expressing emotions through the robot’s eyes and cheeks, requested by the participants of the robot co-creation workshop. As a result of using the module in the robot co-creation workshop, it was found that the eyes and cheeks were selected as the most favorite parts of the robot and tended to be liked.

A body design method was developed to implement a small robot with a soft exterior in a short time. Using this body design method, a robot was developed based on the opinions of the participants in the robot co-design workshop.

In Japan, debilitation due to old age is the second most common reason for people requiring assistance. As standing and sitting movements are indispensable in many areas, making it easier for people to stand up can improve ADL. Therefore, we have developed a wearable device that can be worn on the body to assist standing up movements at all times. In this study, three healthy adult males performed standing and sitting movements and the differences in muscle performance with and without assistance were compared in the rectus femoris, medial vastus, lateral vastus and tibialis anterior muscles. The difference between with and without assist was -1% for tibialis anterior, 2% for vastus medialis, -5% for vastus lateralis, and 22% for rectus femoris in the standing position, and 8% for tibialis anterior, -5% for vastus medialis, -7% for vastus lateralis, and 85% for rectus femoris in the seated position.

Recently, clothing with new functions, such as smart textiles and wearable devices, are gathering attention. Therefore, we propose novel wearable devices with automatic fasten and unfasten functions. In this paper, we developed a device, which was fastened on the wrist and unfastened from the wrist by a simple finger motion. The fastening motion was achieved by a bending mechanism made of a soft material, and a keeping-bending mechanism using a magnet. The unfastening motion was achieved by a release mechanism consisting of a linkage mechanism and a wire. Therefore, the device can be fastened and unfastened by the same simple finger motion. Furthermore, we confirmed the holding force was sufficient for its self-weight. The holding force was increased when it was fastened with the object of larger diameter.

Lower limb exoskeletons have attracted increasing interest and shown considerable potential in assisting human movement. However, traditional exoskeleton designs, which directly attach the actuator to the joint, are challenging to bring real benefits due to the introduced extra mass. In this study, we introduced a bio-inspired Bowden-Cable-Driven (BCD) exoskeleton capable of providing contraction assistance in the plantarflexion direction at the ankle joint while minimizing the additional distal mass introduced by the hardware. A compact and lightweight spool mechanism along with a flexible fabric leg sleeve has been developed to efficiently convert and transmit the torque from the motor to the ankle joint. The corresponding communication framework and feedback-based controller have also been designed to realize fast and precise force control. The actuation performance of the proposed system was characterized through a series of benchtop test experiments. The bandwidth of the system was more than 12 Hz in the force tracking experiment and the force fluctuation under position disturbance was 5%. This study proposed a novel BCD exoskeleton framework for assisting walking and is expected to promote the development of cable-driven exoskeletons.

As a method of preventing knee arthritis caused by age-related decline in muscle strength and bone density in an aging society, this study proposes a non-power-mounted device that reduces the burden on the knee joint during walking by elderly people.

The proposed device consists of a link mechanism and cushioning components, and it is designed to convert the contact force between the sole and the ground into a supporting force during knee joint flexion. In this paper, the optimization method is used to design the dimensions of the mechanism to reduce the disturbance of knee joint motion by the proposed device. The results of the dimensional design were verified using a geometric model to confirm the effectiveness of the proposed device in reducing the inhibition factor on the natural knee motion of the knee joint.

Previous research has analyzed how the assist effect changes due to misalignment of the semi-crouching assist device, and then examined the validity of the analysis by taking measurements. At that time, the problem remains that it was not possible to qualitatively or quantitatively clarify how the support positions of the chest support pad and the thigh differ.

In this study, in order to propose a mechanism to suppress variations in the assist effect, we will elucidate the deviation of the support position of the chest support pad and thigh that occurs when the pulley deviates from the position of the hip joint.

Today, robots are able to perform actions that exceed human physical capabilities. On the other hand, body-enhancing wearable robots have not been able to perform the same functions as limbs. The reason is that the higher the performance of the robot, the heavier it becomes which makes it impossible for a person to carry it. In this study, we propose a body augmented wearable robot, Vlimb, which has both strength and range of motion by using a wire drive and changing the point of articulation to realize limb-like motion. Vlimb has 5-DOFs, a wide range of motion, and can not only manipulate objects, but also successfully lifts 60 kg, the average weight of an adult male.

The objective of this study is to develop metrics for evaluating the safety of wearable robots during unplanned gait termination. In this study, we measured parameters related to the stability of the wearer during unplanned gait termination and explored parameters suitable as metrics. In the experiment, we simulated the occurrence of unplanned gait termination while using the robot in the two conditions with and without assistance that could interfere wearer’s termination motion. Then, parameters were compared between conditions. The results showed that the margin of stability changed due to the influence of the assistance that interfered with gait termination. The time required for gait termination was also confirmed to change due to the influence of the assistance. These parameters, whose changes were confirmed in this study, are expected to contribute to the understanding of the mechanism of changes in gait termination motion due to assistance.

The half-sitting posture and standing up during work are physically demanding. These tasks are often performed outdoors, it is difficult to obtain a power source, the device needs to operate for long periods of time on battery power. The purpose of this study is to develop a wearable device pneumatic that can support standing up and a half-sitting posture. It can reduce the physical burden and realize long-term operation by incorporating a pneumatic regeneration mechanism. We made a small prototype of the concept and conducted experiments to verify its effectiveness in assisting a small humanoid robot to stand up and maintain half-sitting posture. In addition, the effectiveness of the pneumatic regeneration mechanism was also verified through experiments.

Home Oxygen Therapy (HOT) patients often carry portable oxygen apparatuses when going outdoors, which poses physical and psychological burdens. The authors have been developing robotic devices to assist HOT patients during outings, but previous prototypes had issues with user operability. These were attributed to poor operability due to low back-drivability caused by a high reduction ratio drive system, and structural issues in the detection method of cart push force used for the velocity control input. In this study, to address these issues, a new velocity control system using MR fluid clutch mechanism is developed and integrated into the robotic cart drive system. Furthermore, a new prototype incorporating these improvements is designed and built.

In this paper, the effectiveness of a fall risk alert system using a smartphone with a LiDAR sensor is verified based on the behavior of the system in a demonstration experiment at an actual station platform. As a result, we have confirmed that the estimated risk increased when entering the edge of the platform, indicating that this system works in a real environment and can alert fall risk. It was also shown that the alert timing occurs near the tactile paving, which means that the system can provide alerts equivalent to the tactile paving. In the future, we plan to verify the effectiveness of this system in more detail by conducting experiments involving more participants and situations other than the scenarios in this experiment.

Through robot aided rehabilitation, we aim to reduce the burden on therapist in an aging society. Motivation is also an important key to rehabilitation, thus we evaluate motivation using bio-information. In the previous study, we evaluated motivation using electroencephalograms (EEGs). But measuring EEGs does not measure motivation directly. Thus, we use some biometric information as an EMG to improve reliability of evaluating motivation. In the result, there are significant difference between group with additional rewards and group with non-additional rewards. Therefore, we can expect to evaluate motivation by using EMGs.

Supporting the daily lives of frail elderly people with reduced muscle strength is needed. One promising approach is a design tailored to weakening muscles. However, how the shape of the daily environment affects the frail elderly people’s behavior remains unclear. In this study, desks of different shapes were fabricated, and the standing and sitting movements of 85 frail elderly people (65 to 90 year-old) with their hands on those desks were measured. The results confirmed the presence of handrails or handholds significantly increased the force exerted on the desks, but that suppressed the muscle activity contributing to the desk-pushing movement. The handrails and handholds also contributed to shifting the center of gravity of the upper body forward. These analyses suggest that even minor design features can significantly contribute to the postural shift of the upper body of elderly people and assist them in standing and sitting using less muscular force.

To make robot therapy more effective, a tool was developed to improve the therapists’ skills by visualizing their knowledge and beliefs. To implement this tool in a busy nursing home, it is necessary to automate a time-consuming process. We developed a system to automate the extraction of the important actions which are recorded by RGBD sensors during the therapy, and conducted an experiment. As the results, 41 ％ of all important actions could be extracted automatically.

We propose a pair of step-climbing units for manual wheelchair users who have no disabled upper limbs and hopes independent life without caregivers. The unit is a rectangle shape, which comprises a DC motor and a crawler belt. These devices are used combining with a pair of mobile ramps. Each unit moves on each ramp producing propelling force to rear wheels of a wheelchair from both sides. Some of wheelchair users go to work by car. It is important to load the units and the ramps into the car easily. We also make a pair of light and compact prototype ramps using GFRP and aluminum pipes. We discuss the control of the units through wireless communication when descending. Outdoor experimental results when climbing up and down two steps, which height was 19 cm each, confirmed effectiveness of the design.

The purpose of this research is to build a system that a robot arm can quickly and softly move along the shape of the human body, which is often done in physical care situations such as massage, rehabilitation, and palpation. This movement requires real-time visualization of the whole body using a RGB camera with OpenPose, a skeletal detection software, and contact force control by a force sensor. Three types of tracking system which consist of two types of control, speed and contact force control, are used against the still and active body, evaluating them with retention rate and position. As a result, the arm can track body shape adjusting body movement at the low velocity.

This laboratory is engaged in the development and research of robots for practicing caregiving in order to train caregivers efficiently. In this study, we developed a simulation application using a robot for care practice in a virtual space. We aimed to amplify the feeling of dealing with a human being, which has been a problem in previous research, by having a model in the virtual space perform similar movements when the care practice robot is manipulated. As a result, we succeeded in amplifying the sensation that the care practice robot is dealing with a human being by manipulating the robot while viewing a 3D model in the virtual space that is linked to the care practice robot, demonstrating the usefulness of the proposed method.

This study addresses accessibility challenges in the design of construction equipment, aiming to improve usability for individuals with physical disabilities. Despite the construction industry's focus on usability, challenges persist in usability for certain groups. Preliminary research revealed issues with boarding, seating stability, and operation of existing equipment, and the possibility of people with disabilities engaging socially via accessible construction equipment. The study identified individuals with spinal cord injuries Type I (hemiplegia) and amputations as primary target users, with those having Type II (quadriplegia) as secondary targets. Through online surveys, it investigated their specific needs, uncovering a demand for construction equipment in daily activities. Interviews with target users highlighted design requirements including low seat height for easier transfer, minimal third-party support, operability with upper limbs or fingers, and features supporting body stability. The findings defined the target users and detailed design requirements for construction equipment accessible to individuals with physical disabilities.

Parents’ childcare burdens, stress and worries are increasing as toddler grow up. Parents will prepare themselves if they are aware of the unpleasantness of their toddler the next day. This study aims to examine the causal relationship between behaviour and unpleasantness of toddler. Our research group has conducted two-week experiments where the systems for measuring the behaviour of toddler is monitored using the sensor-embedded plushie device, wearable device and emotion recognition in a real-life environment. The result of association analysis shows that the total amount of exercise do not have a large influence on unpleasantness. However, when the total amount of exercise exceeds the average value, the number of unpleasantness in the next day decreases. The cross-correlation analysis using data of sensor-embedded plushie device and the unpleasantness indicates that the peak of unpleasantness due to fatigue caused by playing with sensor-embedded plushie device comes about three hours later.

Although sensing technology has advanced in recent years, little attention has been paid to understanding the relationship between the benefits and exposure of personal information using sensors. This lack of understanding is a major factor hindering introducing sensing systems in daily life. To deal with ELSI related to sensing technology, this study proposes a Willingness to Expose (WTE) method to clarify the relationship between the amount of information exposure and the benefits of sensing services. The authors conducted a questionnaire survey for 500 people using the proposed WTE. The results of the WTE survey show that the more attractive the service is, the higher the WTE is (p<0.0001). The survey also shows that the WTEs significantly differ depending on where the sensing systems are installed (p<0.0001).

Unintentional injuries are the leading cause of childhood death and falls in particular are the most common mechanism of death and serious injuries. In this study, first, in order to understand the environment in which falls actually occurred, the authors visited 27 houses and analyzed the characteristics of the sites where falls occurred by measuring three-dimensional shape data using a scanner. Next, based on the characteristics of the sites, experiments were conducted with 37 infants aged 13 to 45 month-old on climbing capability and moving speed that have not yet been elucidated conventionally. Finally, based on the behavioral characteristics revealed by the measurement results, a fall simulator was built. This simulator allows us to predict whether or not fall accidents occur, the routes that lead to the accidents, and the approximate time until the accidents occurs, depending on the shape data on the target environment, the developmental stage of the target child and the child’s location.

Contacts play an important role in natural robot interaction with humans and environments. As humans generate such motions with ease in their daily lives, understanding their motions can serve to improve robots’ capacity. However, access to human motion data including contacts is still limited. In this study, we propose a framework for measuring human motions involving surface contacts by collecting data from distributed tactile sensors and motion capture systems simultaneously. The contact information is aligned on the human body through position-orientation registration and unified with synchronized body motion data. Some examples of measured contact motion are presented to demonstrate the functionality of the proposed framework.

In Japan, the proportion of elderly people living alone is increasing, and there is a concern that in emergencies such as accidents, they may be unable to contact external help. Therefore, this study aims to develop a fall detection system for monitoring elderly individuals living alone. Using a depth camera to measure body acceleration, sudden accelerations during falls were confirmed. Furthermore, we created an inverted pendulum model in a dynamics simulator, conducted simulations of acceleration characteristics, and compared them with the measured acceleration during falls. As a result, it is considered that the inverted pendulum model is useful as training data for fall detection.

We focus on the torso and shoulders and analyze the movements of these joints during specific tasks in order to generate humanoid movements that take advantage of redundant joints. In addition to grasping an object and shifting the center of gravity, we selected a task that requires rotation of the torso to carry a package between shelves for analysis. For this movement, the joints whose joint torque values changed significantly when the mass of the load to be transported was changed to a larger one were considered to be the joints that contribute significantly to the load-transportation movement, and were the target of the postural analysis. The joint torques of 20 joints in the whole body and a total of four joint angles (torso flexion extension, torso lateral flexion, torso rotation, and shoulder flexion abduction) were measured for the evaluation.