Tulip bulbs are the main production item of flowers in Toyama Prefecture, and are also prefectural flowers. Among the tulip bulb production processes, the disease detection and extraction processes rely on visual inspection by skilled farmers and are not mechanized. Therefore, we tried to detect disease using images of growing tulips in order to contribute to the mechanization of this process. It was found that higher discrimination accuracy could be obtained by creating discriminators for each age of the tulips.

Prepared foods and/or convenience store lunches have a short product cycle and are made up of various foodstuffs, so their manufactures have to rely on manual labors. Labor shortage has become apparent so that the robotic and automation system would be beneficial to save labor shortage and reduce tasks. The robot system is required to identify various foodstuffs instantly, and to pick up and transport them to a predetermined position of the container. This study reports on the development of various foodstuffs selection from RGB-D images without changing the sensor configuration, and picks up and transports the target foodstuffs using a robot hand suitable for each foodstuff.

Developing an autopilot carrier for orchard on sloping land is effective for labor-saving in the operations on harvesting and transportation. An autopilot carrier’s positional information required for automatic driving is obtained from GNSS. However, the positioning accuracy decreases because the fruit trees blocks the GNSS signal. In this paper, we evaluate the effect of fruit trees on RTK-GNSS positioning accuracy.

Recently, the labor shortage is becoming more serious because of decrease and aging of farmers in Japan. As a solution，therefore, the automation by agricultural robots is trying to expand the scale of management and labor saving. In order to automate harvesting, we use semantic segmentation, a method of deep learnig, to discriminate onion, soil, and others in pixel units. Then, the height of only the soil part detected using a depth camera is measured. Harvesting with high precision is performed by using the information to automate height control of the harvesting part. The usefulness is confirmed by some experiments.

In the spinach automatic harvester which has been developed by the authors, pitch angle fluctuation sometimes deteriorates spinach harvesting accuracy. It is caused by resistance force from soil, and previous studies have shown that the resistance force from soil can be reduced by considering the moving speed of the root cutting blade into soil. In this paper, we examined the improved path which suppresses pitch angle fluctuation by taking into account the moving speed of the root cutting blade into soil.

In the use of robots in agriculture, small robots have the advantage of low invasiveness to crops and fields, but on the other hand, there is a problem that they are difficult to run due to environmental factors such as terrain. In order to solve this problem, we prototyped a small paddy field robot that has a mechanism in which wheels follow the bottom of the water according to the water depth. Effectiveness of the proposed method is confirmed by water tank experiments.

The purpose of this study is to develop a power assist device with crawler moving mechanism in order to reduce the burden on the grape growers. The power assist device is mounted to a mobile robot by a flexible arm and consisted a linkage mechanism in order to use a single actuator. The mobile robot consists of two tracked vehicles in order to work in the vineyard. The effectiveness of the proposed power assist device is verified by the result of the experiments.

This paper first describes the outline of the tie-stall milking robot. We analyze the changes in the working style and workload of the farmers before and after the introduction of the robot using video images and activity measurement. Finally, we describe the effect of reducing the burden of introducing a robot and future issues.

This paper describes a method of dataset generation to realize cucumber recognition for harvesting robots by deep learning. To recognize cucumber fruits, conventional image processing methods are not very effective because cucumber fruits, leaves, and stems have same color, while deep learning is useful because it can automatically learn features to recognize them. Deep learning, however, requires a large amount of data for training, and there are no datasets for the cucumber greenhouse. Manual annotation consumes time and cost. Therefore, we developed a method to generate a large dataset automatically using computer graphics. We generated a dataset by creating and rendering 3D models of cucumber plants with their actual parameters measured in the greenhouse. Using this method, we made a cucumber dataset for semantic segmentation, object detection, and instance segmentation.

The burden of agricultural workers has increased to weed grass on levees in rice fields due to the aging of agricultural workers in Japan. To decrease workers' burden, weed cutter robots with the function of changing the camber angle system have been developed in our previous study. However, the camber angle sideslip suppression effect is not enough to suppress sideslip when the inclination is 40 degrees or more. To solve this problem, we propose the sideslip suppression method using the camber angle and steering angle. From experiments, it was confirmed that the proposed mechanism was effective at an inclination of 40 degrees.

In recent years, the aging of agricultural workers and the reduction in work force is progressing. Especially in farm work, the management of levee is very burdensome for workers. Conventionally, a brush cutter has been used for levee management, but this task is very dangerous for workers. Therefore, various robots have been developed for levee management. In the development of robots, it is important to suppress sideslip and miniaturize to improve working efficiency. In this study, we propose novel traveling method that can suppress sideslip and improve working efficiency by rotate the wheel alternately left and right to sway the body.

Agricultural workers have required automatic harvesters due to aging of the Japanese people recently. Spinach is one of the difficult vegetables to harvest automatically. Therefore, Shinshu University proposed a harvesting method called "passive handling". The previous works revealed that the difference in root cutting blade path resulted in a difference in harvest success rate. The compared paths were horizontal path and arc path, and arc path gave better results. The difference in performance between the two paths has remained unclear. This study clarified the reason why the arc path is good from the viewpoint of the interaction between the root cutting edge and the soil.

In this paper, we propose a robust algorithm to detect landmarks in a vineyard field. The algorithm only uses image data from inexpensive web-camera sensor to accurately detect pillars supporting the grape plants in the presence of noise. Robust detection of these landmarks is important for the autonomous task execution by a robot in vineyards. The detected pillars are also used for accurate localization and mapping. Moreover, based on accurate pillar detection, semantic data logging in vineyard becomes feasible. Traditionally, expensive GPS and Lidar sensors are used for landmark detection. The proposed method can robustly detect the supporting pillars or the vineyard using inexpensive cameras on robot.

This paper introduces an artificial marker system for underwater environments using acoustic camera. The system is composed of detection of the designed square markers, ID identification and six-degree-of-freedom (6DoF) pose estimation. The proposed system can be applied to tasks such as underwater positioning, underwater vehicle navigation and augmented reality (AR).

We use Remotely Operated Vehicle (ROV) for capture sea urchin. Sometimes ROV hard to catch sea urchin on a specific terrain. ROV used sea urchin capture unit, but sometimes it didn’t work because the end of unit is short and can’t control it directly. So, we developed Sub System for ROV. It can change the pitch angle up and down with two thrusters. And It have several sensors, will help our work. This paper describes the composition of Sub System and experiment results using it in the real work.

Valuable seabed mineral resources such as rare earth elements are deposited 2–3 meters into the deep sea floor, and sampling and analyzing the seafloor is necessary to clarify how they are generated and distributed. We have developed a seafloor robotic explorer that can excavate horizontally and collect samples of rare earth elements, which may enable wide-area exploration as multiple robots could be simultaneously implemented to autonomously search beneath the seabed. The seafloor robotic explorer may excavate sediments such as clay and sand before reaching the rare earth layer. In previous studies, it has been discovered that the rotation speed, penetration force, and characteristics of the sediment affected the excavation resistance. Therefore, we assumed that controlling the rotation speed and penetration force would enable an efficient excavation according to the ground characteristics. Therefore, in this paper, we examined the effect of penetration force on drilling efficiency using both dry and underwater silica sand ground with different ground characteristics.

In this study, a small model underwater glider that can measure the position and orientation during gliding is developed. To understand the properties of the robot in the water is important, e.g., the glide ratio to reduce the energy consumption. The glide ratio depends on the shape of the wings and body. We are developing the wing simulator to realize the simulator for whole motion of underwater robot. For checking the accuracy of the simulation, the experimental results are compared with that of the simulation. The position and orientation of the model underwater glider are measured by a tilt sensor and a camera, respectively. Finally, the gliding experiments are conducted to evaluate the measured position and orientation.

In this paper，we propose a Center-of-Gravity(CoG) movement device using the differential screw drive mechanism for glider type small underwater robots．The differential screw drive mechanism consists of a right-hand and left-hand screw-nut mechanism．The rotational and translational movement of nut is achieved by the differential motion of two screws, and the position of the CoG is changed according to the movement of a weight．Therefore，the CoG movement device can control the posture of the robot．Double-nuts and springs are embedded to the CoG movement device to remove the backlash．We evaluated the effect of the backlash removal mechanism by experiments and confirmed its effectiveness．

The depth of water is important information for the ship navigator when the ship enters the port. However, the depth is not always the same as the chart due to tides and sedimentation. This study aims to develop a system that measures the water depth by Underwater Vehicle and provides the information to the navigator. In addition, the sounding by the developed underwater robot was carried out, and the water depth map of pond was successfully created. This paper describes the hardware configuration and system design of the developed Underwater Vehicle, and the experimental results using the Underwater Vehicle. The tide calculation is summarized in the discussion.

In this paper, we compared the sealing performance and friction force of different shaped PTFE rings. PTFE ring is a seal ring used for buoyancy control device. The PTFE ring seals the PTFE slide shaft. When the friction between the PTFE ring and the PTFE slide shaft is reduced, the operating efficiency of the buoyancy control device is increased. PTFE rings with multiple lips and a slipper seals are compared.Experimental results showed that the PTFE ring with single lip had more than 1 [MPa] sealing performance and the lowest friction among the comparisons.

The current method of scallop fishing is mainly Scallop Dredge. When sand at sea bottom enters a scallop during harvesting, none of the scallops can be sold for raw. Additionally, Scallop Dredge harvests a scallop with marine sources such as other shells. To avoid these problems, this study proposes a new harvesting method using a robot arm. The new method can catch the only scallop without entering sand in a scallop. the new method needs a gripper to attach a gripper to a robot arm. This study focuses on the phenomenon of adsorption by negative pressure and tries to develop a negative pressure adsorption gripper for scallop harvesting. an experiment gripping a scallop with the gripper of this study was successful.

In future underwater development, it is considered that many tasks will be achieved by cooperative motions of underwater robots. In this paper, as a first step in the development of a cooperative control method for multiple underwater robots, we report simulation results of object manipulation by two robots, each of which has a position control system.

The main underwater position measurement system is the measurement method using sound waves (e.g., SONAR). However this method is generally low accuracy and very expensive. The purpose of this study is to develop a position measurement system with higher accuracy than conventional underwater position measurement systems. In this study, we proposed a position measurement system using optical sensor as a new position measurement system in the underwater, and an experimental system was created to simulate the proposed position measurement system. The problem of the proposed position measurement system is the tracking performance of the tracker. Therefore, the tracking performance of the tracker which is a part of the proposed position measurement system, is evaluated by the created experimental system.

Marine debris causes deleterious effects on the environment and economy. It is still a difficult task to efficiently collect marine debris floating on the sea surface. We previously proposed a semi-automatic collection system of marine debris using an underwater robot. However, the success rate is low due to the complicated operations required to collect marine debris. In this research, we propose to use a suction mechanism to simplify the operation of collecting marine debris that enables the underwater robot to collect just by approaching it. The experiments verify the effectiveness of the proposed system with respect to the success rate and time of collection.

We have developed Multidirectional INtuitive Aquatic MObility (MINAMO) that is a personal vehicle on water. In this paper, we developed MiniMO whose size is approximately a wash tub to perform autonomous self-position control. We experimentally show that MiniMO can autonomously move on water by means of Pozyx that is a positioning device using radio wave to locate.

Seabed mineral resources such as rare earth elements are widely and shallowly distributed in the deep sea floor 2-3m. Seafloor sampling and analysis are needed to determine their distribution. Vertical drilling with a drilling vessel is effective for depth sampling. However, it requires enormous cost and limits the sampling range to the pipe diameter. Therefore, we are developing a seafloor exploration robot that can excavate horizontally and collect rare earth samples. In this paper, we report the result of field test in an actual sea floor using an excavating robot with water proof finish.

In this research, we improved the motion performance of a portable underwater robot that performs underwater work on flat or curved surfaces by an adsorption force generated from thrusters with negative pressure effect plates. The developed robot has passive joints so that the angles of the thrusters can be changed according to the curvature of walls. We also conducted a preliminary experiment to investigate the force generated by the negative pressure effect plate.

In recent years, the research related to autonomous and remotely-controlled ships has become the major topic in the maritime field. The 99th Maritime Safety Committee (MSC) of the International Maritime Organization (IMO), many delegates declared that ships with remote control from shore or with partially autonomous functions would be widely used for a long time before the realization of unmanned operation. In our laboratory, a prototype of the remote navigation system has already been developed. Currently, we are researching support systems for ship operators. In this paper, we consider that it is necessary to confirm execution of maneuvering command for the remote maneuvering operator not only with the measuring instrument but also with the actual measurement. As a preliminary experiment, the results of detecting the characteristics of the operation sound in each ship maneuvering instruction are shown.

As well known, the Arctic Ocean are covered with vast and thick sea ice, it is difficult to grasp the state below sea ice from above sea ice. Analyzing the behavior of sea ice is directly related to the understanding of the Arctic region, and leads to a global understanding of the Earth. For the observation of sea ice above the sea surface, some methods using an airplane or a satellite has been established, however this method cannot observe sea ice below the sea surface. In order to observe sea ice in the Arctic Ocean, observation of sea ice from underwater is essential. For this reason, JAMSTEC is now developing an autonomous underwater drone called as COMAI that can navigate under the Arctic sea ice and continuously observe the shape and seawater condition of the sea ice. In this paper, the outline of COMAI is explained, and its kinetic performance is shown from the tank test results.

In recent year, the human collaboration robot (HCRs) have attracted, and HCRs with function of contact detection are commercially sold. However, safety measures for end effectors on HCRs may be not sufficient. In this paper, we propose a safety light curtain using ToF sensors for safety measures of end effector on collaboration robot. The light curtain mounted on the top of the end effector can detect the object around the end effector without contact. Thus, the HCRs with end effector can avoid an unexpected collision using this sensor. Therefore, the safety of cooperative robots can be improved.

We calculated the driving conditions based on the upper limit value of the falling moment derived from the static stability performance. And we conducted a dynamic stability test to understand the overturning limit performance of the autonomous mobile guide robot during running and confirmed any behavior. As a result, it was found that it is desirable to focus on the portion where the inertial force acts, and that a dangerous spin phenomenon occurs when turning without turning over.

Our previous study proposed a nursing assistant system using deep learning and robot technology for patient safety. However, there are characteristic risks associated with artificial intelligence (AI) in applications such as nursing assistant system. In this paper, we present safety applications related to AI in fields where humans and robots coexist, especially when applying deep learning to the control of assistive autonomous mobile robots. First, we systematically extracted risks, risk manifestation phases, risk factors, and cases based on the results of patient simulation experiments performed on the campus. Next, we proposed risk reduction measures from the viewpoint of safety control. Finally, as a case study, we quantitatively evaluated the effects of risk reduction measures and confirmed its effectiveness of reducing the error rate during learning by about 30%. Furthermore, this method can be applied not only to hospital facilities but also to residential areas including nursing homes where the most incidents occur.

The measurement of internal stress and strain distribution inside the soft material, which is demanded in the field of skin contact safety, are hard to be obtained nor estimated using conventional interactive force sensors except for limited situations. In this study, a shear strain sensor system based on the distributed built-in piezoelectric Polyvinylidene Fluoride (PVDF) polymer films is developed for measuring the shear strain field inside soft material and evaluated. The sensing principle is modelled by relating the local shear strain of substrate and the charge generation induced by piezoelectricity in the in-plane mode. The validity of this sensor system was tested by measuring the three-dimensional shear strain distribution of an artificial human tissue under simple shear deformation. An applicable consistency calibration method is also considered for this kind of multi-sensor system. The detected shear strain distributions by this sensor system show a good agreement with the numerical simulation of the substrate. The proposed sensor system using distributed built-in PVDF films is confirmed to sensitively detect the distribution of the shear strain inside soft material.

Nowadays, with the aging of society, the spread of safety requirements for personal care robots that assist people's lives autonomously is desired. As personal care robots become widespread, it is expected that someone who physical weakness such as children, isolated from industrial robots, coming into contact with the robots will increase. Though, the safety standards for personal care robots require to give consideration to safety for children, they do not specify specific methods of the consideration. Therefore, it is necessary to clarify necessary items when considering child safety for the personal care robots. In this study, we will clarify the items necessary when considering safety for children and inestigate experimental methods for evaluating safety considering children in contact.

The safety measures of the omni-directional autonomous mobile robots for work support have been subjects. Authors developed the speed monitor module which can reduce a collision risk with human and other safety-related objects by adding to a robot's control system. The developed speed monitor module calculates a robot's running speed by detecting the number of rotations of four wheels of the omni-directional mobile robots, and outputs the signal for protection area selection of a safety laser scanner according to speed. The performance level of the module according to ISO 13849-1 was evaluated, and also the improvement of a performance level was considered.

The labor shortages are becoming more serious due to the declining population in Japan, and as a result, the use of robots is expected to expand in all industrial fields, including the manufacturing industry and the service industry. In particular, cooperative robots that coexist with humans, unlike conventional robots that have been separated by fences from human to ensure safety, are beginning to spread. In such situation, it is thought that there are risks that cannot be dealt with by conventional concept of safety of machinery. Therefore, in this study, we show with a new model that safety aspects of human-machine coexistence systems should be considered from a dynamic aspect and clarify the countermeasures.

The inherently safe design, which is prioritized by the international safety standard ISO12100, requires assurance that a machine will not inherently output more than humans can withstand. MR fluid is a functional fluid that can control the phase change from liquid to solid by magnetic field, and is highly responsive. At present, there are many studies using MR fluid, and the number of research reports is increasing. However, no control system that actively utilizes the characteristics of MR fluids has been implemented. Here, a control system that actively utilizes the characteristics of MR fluid and transmits the available energy by the MR fluid when necessary is newly developed and its control performance is verified. Since large amounts of energy are not transmitted by slipping of the MR fluid, it is possible to suppress output to within safe limits for humans.

In recent years, digital transformation has attracted attention. In such a situation, since the exchange of data between manufacturing equipment and control equipment increases with the adoption of IoT, there is a possibility of security problems occurred by duplication of manufacturing robots in factories. Therefore, the management method of the robot is not based on data like QR code, but is studied by attaching a unique physical pattern. In this study, we focused on the fact that a tuning pattern can be obtained by using two soft materials with different curing times, and produced a Physically Unclonable Artificial Skin (PUAS) with various surface patterns by changing the mixing ratio. As a result, the characteristics of the patterns by mixing ratio were found.

Risk assessment and risk reduction are the keys for safety robots. Even on demonstrative experiments, risk assessment is required for safety robots. In the case of demonstrative experiments, a series of robots are evaluated at various test fields, and those robots may have different type of optional functions which fit to the test fields. From the point of safety robots, each type robot shall be evaluated through risk assessment and reduce risks so as not to harm any human being and property. This paper shows the practical way to reduce the cost of risk assessments for a series of robots. The method may applicable to manufacturing process of a series of robots.

We experimentally verify the risk of small UAV propellers. We are studying the necessary safety equipment for flight engaged person. In this paper, we focused on the eyes and performed an experiment to confirm the effectiveness of safety glasses. As a result of the experiment, it was confirmed that general glasses (glasses that do not comply with the standard indicating impact resistance) are insufficient as protective equipment.

Because of the aging society with a declining birthrate, labor shortage for people who need rehabilitation is one of serious problems in Japan. On the rehabilitation site, physical therapists care patients one‐to‐one and determinate their conditions based on experience and memory. Therefore, we need to develop systems which evaluate automatically and quantitatively without expensive devices such as force plate. Balance ability is an important factor of motion control function and is related to basic motions which affect Quality of Life: QOL such as standing and walking. In this study, we try to estimate instability of human body balance based on centroidal angular momentum instead of using force plate and to evaluate it automatically by making a balance game using motion capture．

In this study, we investigated a method for operation adjustment of myoelectric prosthetic hand by voice recognition. Using a robot hand that can be controlled by EMG signals, we developed a prototype control adjustment module by voice recognition. This paper reports the results of experiments using the prototype.

In previous work, we developed a shear-force sensitive sheet (SSS) that detects the shear displacement and shear stress that occurs between humans and machines. In this paper, we improved the SSS by adding protective sheet (P-SSS). That helps to start recording at any time. According to the experimental result, significant differences were found between normalized area and average line length of P-SSS and those of SSS.

Children with impaired motor function may be difficult to move on their own. Our research group produces a prototype of an omni-directional mobile electric wheelchair, that has high degree of freedom and can move close to the movement of healthy people, and offers it to children with disabilities. We can provide the easy and intuitive operation by the proposing omnidirectional mobile electric wheelchair. However, children with poor control skills may not be able to operate satisfactorily, and may collide with obstacles, and may not be able to escape from the collision state.

In this research, we considered “collision experience” as “learning opportunity” and worked on the development of an intelligent bumper that generates the safe collision and encourages the learning for escapes.

In home-visit rehabilitation, a quantitative and easy-to-use system to evaluate motor function of the elderly is in demand. After the interview with physical therapists of asking the requirements for the system, we selected three functions from ten candidates for the development of the application. The application measured skeleton data from the users with a compact RGB-D sensor and a skeleton tracking middleware. The results of data processing were displayed on the computer screen. One application showed CoG loci in three 2-D projection charts, second one printed foot positions of a step movement, and third one presented the CoG loci as 3-D figure. Physical therapists evaluated the developed applications as the features of the motion was recognized easier and was helpful for evaluation of the motor function.

In this research, we are studying the pet robot used to improve dementia. This pet robot is focus on animal therapy and reminiscence therapy. It is a robot that looks like an infant, and is characterized by being cared for by the target person and extremely lightweight. However, our pet robot had several problems. This robot’s moving was too simple and boring. For that reason, we change the structure of the robot. New robot is lightweight maintain and diversifies its movements. Therefore, the adoption of wire drive allows the movement of multiple joints to be reproduced with a single motor, thereby achieving weight reduction.

Sustainable use of home healthcare devices requires high usability. For example, light weight, small size, simple operation method, and the like can be cited. In this study, we focus on shoulder massage equipment that is large and tends to have limited posture. In this paper, we propose a wearable massage wear using a soft actuator made of shrinkable film, and describe its principle and structure. Furthermore, two types of experiments were performed as evaluation experiments using them. We evaluated the fitting function that imitated the human body by measuring the distribution pressure on the human body as necessary. Next, the usefulness as a massage device was discussed by measuring the blood oxygen level.

In this study, we evaluated the usability of a cycling wheelchair by changing the seat position. A cycling wheelchair is one of the system that lower-limb disabled patients are able to move easily. However, when the patient uses the cycling wheelchair in the daily life, it has some structural problems. In order to resolve the problems, we focus on the position of the seat. We evaluate the influence that between user ’s pedaling motion and the position of the seat. For analyzing, two links model is used to estimate hip and knee joint torques. Finally, we conducted the experiment to confirm the effect of the seat position.

This paper proposes a training system for a multimodal comprehensive care methodology for dementia patients named Humanitude. Humanitude has been attracting much attention as an effective care technique for dementia patients. Humanitude consists of four main techniques, namely, eye contact, verbal communication, touch, and standing up, and more than 150 care elements. Learning Humanitude thus requires much time. To train Humanitude effectively, we develop a training systemthat realizes sensing and interaction simultaneously by combining areal entity and augmented reality technology. To imitate the interaction between a patient and a caregiver, we superimpose a three-dimensional CG model of a patient’s face onto the head of a soft doll using augmented reality technology. Touch information such as position and pressure is sensed using a whole body wearable tactile sensor. The effectiveness of the proposed training system is evaluated in public lectures.

In this study, we aim to quantify care skills in the dementia care technique called "Humanitude". Humanitude has been attracting much attention as a gentle and effective care technique for a dementia patient in recent years. The developed wearable whole body tactile sensor aims to quantify the "touch" skill among the four representative skills in Humanitude, which are gaze, touch, talk, and stand up. In addition, we conducted a demonstration experiment using a wearable whole body tactile sensor to quantify nursing care by general caregivers and a Humanitude expert.