In this paper, we propose an improved distance measurement method using blur for 3D measurement. Stereo disparity is often used for 3D measurement, but when a target object is too close to the sensor, it cannot be obtained properly. To solve this problem, we develop a compact range image sensor capable of close-range measurement focusing on the amount of blur using a multi-slit laser and a camera.

Currently, automated driving has reached level 3, and research and development is progressing toward level 4. Driver and passenger safety must be guaranteed in automated driving, there are various causes of accidents, but the accident rate in bad weather is about five times higher than in fair weather. Focusing on this, AI is used to recognize obstacles in bad weather. Specifically, first, the features of the vehicle are extracted. Next, the original image is fog-processed to create fog image data. Let the AI learn it. In addition, in terms of cost reduction, we aimed at obstacle recognition with a low-cost monocular camera.

While there is a growing emphasis on improving QOL of patients with neurological diseases such as ALS and SMA, the sport of boccia, which can be played by people with relatively severe disabilities, is attracting attention. In this study, we developed a system that measures the distance between balls from the jack-ball and displays the results of victory or defeat from each distance value, with the aim of assisting patients with neurological intractable diseases to compete. In this system, each ball was detected by a labeling process and a circular approximation of the Hough transform, and the distance between balls was calculated from the distance values to the balls by the RGB-D camera. In the verification experiment, the system's recognition accuracy and distance were investigated. The results showed that the average ball recognition rate was 98.4% and the measurement distance error was 4.4%, confirming the effectiveness of the system.

This paper describes a 3D mapping method with object recognition result by using a wheeled mobile robot. The object recognition is realized by using a deep learning neural network called PointPillars. By comparing the recognition results between two continuous frames, the object ID can be matched and for those which are recognized in both frames, and ID will be changed for the others. The recognized objects with determined ID will be reflected to the 3D map.

In this report, we propose a flexible balloon sensor for detecting the foot-ground interaction of four-legged robots. The sensor consists of a balloon made from silicone rubber, a tube, and a pressure sensor, which can be installed on the robot’s toes. When the toe touches the ground, the balloon deformed, then the air pressure inside the balloon increases, which can be measured by a pressure sensor. We fabricated a balloon using molds fabricated by a 3D printer. We conducted an experiment to evaluate its effectiveness and the delay time. We built a jig with a photoelectric sensor for detecting the moment when the sensor touches the ground. In addition, we also conducted an experiment to investigate the effect of length of the tube connecting the balloon to the pressure sensor on the delay.

This study analyzes the recognition process of examining the texture of fine particles through tactile sensing using human fingers and aims at establishing the structure of texture recognition. Our previous study confirmed the occurrence of hysteresis in tactile sensation in the mid-teens. This study increased the number of subjects and reproduced the mid-teens hysteresis phenomenon occurring in tactile sensation. The results show that the hysteresis phenomenon can be reproduced in mid-teens tactile sensation, as in our previous study. Furthermore, this study investigated the effect of the delay in sample presentation to the subject on hysteresis. The results show that the hysteresis phenomenon changes as the sample presentation is delayed.

Most sports training in Virtual Reality uses only eyesight and hearing, not touch. So we considered that we gave haptics to VR sports. Among them, we focused on the catching movement, which is the basis of skills such as baseball and softball. To improve the reality of catching a ball, we proposed to present the sensation of catching a ball to the palm using traction force and electrical skin stimulation. As a result, it was partially confirmed that the sensation in catching was improved by the electrical skin stimulation compared with the traction force alone, but the difference from real catch was remained.

This paper describes the development of a high-sensitivity, high-dynamic-range barometric tactile sensor that can be mounted on a cylindrical shape and can detect contact in all directions within 360 degrees, for use in robot arms and snake-like robots. As an evaluation experiment using the prototype sensor, a load was applied via a flexible sheet wrapped around the sensor by 180 degrees, and a linear response of the sensor output value was confirmed for tensile loads from 3 gf to 6000 gf.

3D food printing for shaping food is a new technology that is gaining attention in the food industry and food science field. And 3D food printers have contributed to the advancement of food texture research. The application of food printing is expected to lead to the development of new ways of using it and expanding opportunities. By selecting a 3D food printer that can precisely control the manufacturing process, the formation of foods with complex food textures will be implemented. In this study, we examined the food material composition for 3D food printing, 3D food shaping, and post-processing for developing complex oral tactile.

Evolution of XR technology and Metaverse in recently, flexible piezoelectric devices are required for virtual reality and communication robots. We are developing pre-charge and MEMS tactile device. And we extended to multi-points actuation. This device is very thin and confirmed the vibration.

This paper developed a wearable tactile device with a finger joint angle measurement function and a pressure presentation function to realize an interactive handshake remotely. The measurement experiments conducted in conjunction with the development of the device control system showed that the pressure acting on the palm of the hand was greatly influenced by the other party, while the pressure acting on the fingers was greatly influenced by the self.

Predicting the grasp stability before lifting an object, i.e. whether a grasped object will move with respect to the gripper, gives more time to modify unstable grasps compared to after-lift slip detection. Recently, deep learning relying on visual and tactile information becomes increasingly popular. However, how to combine visual and tactile data effectively is still under research. In this paper, we propose to fuse visual and tactile data by introducing self-attention (SA) mechanisms for predicting grasp stability. In our experiments, we use two uSkin tactile sensors and one Spresense camera sensor. A past image of the object, not collected immediately before or during grasping, is used, as it might be more readily available. Dataset collection is done by grasping and lifting 35 daily objects 1050 times in total with various forces and grasping positions. As a result, the predicted accuracy improves over 2.89% compared to previous visual-tactile fusion research.

According to the method of support for the visually impaired, as for orientation, the support method of guiding them to the destination has some problems, such as not being suitable for changing the goal or deciding the direction themselves. Therefore, we aim to develop a system that tells them information of the left and right sides without hindering their walking and losing information about sound and haptic. In this study, we proposed to present tactile at the sole. And we investigated the perception of changes in vibration when the pace is changed. As a result, there was a little delay in the perception of vibration changes regardless of pace.

This paper proposes a flexible teethed belt (FTB) that can bend multi-direction to coexist the flexibility of the wire drive and the transmitting performance of the teethed belt drive. The FTB consists of a looped mandrel and rubber teeth, which can be fabricated by rubber molding. To apply the FTB to a remote drive with a rotation axis orthogonal to the input/output axis, we designed a belt path that keeps a belt length if the orthogonal axis rotates. Through experiments with the constant angle of the orthogonal axis, we verified that the FTB shows almost the same transmitting performance independent of the angle of the orthogonal axis.

In order to achieve both safety and speed in cooperative manipulators, it is important to reduce the weight of the moving parts. Coupled tendon-driven manipulators can reduce the weight of moving parts by using a redundant number of wires in relation to the number of joints and integrating the motor into the root link. On the other hand, the joint structure tends to be large and heavy because a large-diameter pulley is required every time a wire passes through a joint. In this study, we propose a Passive 3D Wire Alignment System as a lightweight and low-friction mechanism that enables the transmission of wires across multiple joints. Furthermore, a lightweight 7-DoF manipulator using this mechanism is fabricated, and the usefulness of the mechanism is verified.

In this study, a spring tightening and loosening device is investigated for the development of a robot prosthesis with a variable stiffness function. This mechanism has a spiral bolt-like shape and grips a spring as a cover. We have confirmed that the proposed mechanism works and is effective, and we are aiming to realize an electrically driven mechanism in the future. This paper describes the development of an effective power transmission mechanism for the proposed structure. In this study, we proposed and fabricated a mechanism based on a superimposed structure of square columns that can accommodate variable lengths of the transmission section. Although some accuracy issues were identified, variable spring compression could be achieved, demonstrating the effectiveness of the proposed mechanism.

We are developing clustered convex type telescopic manipulators. Compared with the conventional linier actuators, our convex type manipulator can make long reach and are very light. They are also compact when they are reeled up. Even for these advantages, our previous models have a risk of bending because the quasi-pipe like telescopic structure was based two tapes with parallel bending directions. In this report, we will describe the new manipulator type-K2, which employs crisscross-conformation of tapes for stabilizing of the quasi-pipe conformation with four tapes with cross bending directions, to improve the total performance. The tape driving mechanism has specially designed structure for the tandem conformation, for quasi-4 tape operation.

Due to the social background of declining birthrates and an aging society, object transport requires automation and the introduction of robots. This study focuses on sphere transport as a means of efficient object transport by robots. This paper uses a spherical motion measurement system to evaluate the forward kinematics of sphere transportation. The results show the validity of roller-sphere forward kinematics in three patterns of roller arrangement.

Omnidirectional crawler mechanisms have been proposed to improve the mobility of conventional crawler mechanism, but they have problems with twist of crawler belt and travelling ability. In this paper, a twin belt crawler mechanism with circular cross section is presented to improve twist of crawler belt and travelling ability.

In this paper, we examine a method to make it easier for robots to move on soft ground by hardening the ground. After examining several methods of hardening the ground, we came to the conclusion that calcination of grain flour dissolved in water is suitable.

Home Oxygen Therapy (HOT) patients must take the conveyance cart to carry oxygen equipment when they go out. However, the equipment is heavy enough to discourage the patients from going out. Therefore, in this study, we attempt to develop a mobile robotic conveyance cart to support HOT patient's going out by applying the torque convertor using MR fluid. In this paper we propose a new control system and show the effectiveness of it by numerical simulation.

In this study, we developed a shopping assistance robot system. First, we attempted to create an algorithm to find an efficient route to visit multiple sales floors of a product. This algorithm is based on the solution of the Traveling Salesman Problem. The Traveling Salesman Problem is the problem of finding the shortest path that visits all given cities one at a time. By combining the 2-opt method, which is one of the methods for solving the traveling salesman problem, and the A* algorithm, which is a method for finding the shortest path between two points, we were able to create a route that efficiently visits the sales points of multiple products.

Robots with lugged wheel demonstrate excellent traversing performance on sandy terrain. However, when moving on the rigid surface, large vertical vibrations are generated, resulting in energy loss and mechanical failure. In this study, we have developed a wheel with asymmetric elastic lugs which effectively reduces running vibration, while maintaining the high mobility on the sandy terrain.

A mobile robot that can transition during travel between two-wheel modes with two wheels arranged in parallel on the left and right, and front and rear two-wheel modes with wheels arranged in a straight line is proposed. To compensate for balance during the transition and make experimentation easier, we have attached an inertial rotor to the robot. The control method and results of this inertia rotor are reported.

We are developing a crawler type robot that automatically inspects outdoor equipment. This crawler robot patrols the factory grounds to read meters, check pipe temperatures and leaks, and measure the sound of pump motors. Set the travel route and inspection points while operating the robot from the remote-control UI. We built a system that automatically patrols a set route and records images of inspection points.

Autonomous mobile robots are being utilized in various locations such as factories, residential and commercial buildings, and farmlands, and the market for autonomous mobile robots is expanding. Safety is a critical requirement for autonomous mobile robots to avoid collisions with surrounding objects. In this paper, we propose a control scheme based on the distance between a two-wheeled vehicle robot and an obstacle for collision avoidance. We implement a control barrier function to assist in collision avoidance on an actual machine. Additionally, we place the actual machine on a vibration machine and run it under conditions where artificial irregular disturbances are generated. By using a control law based on a control barrier function for probabilistic systems, we demonstrate the safety that can be guaranteed against irregular disturbances.

Wheeled mobile robots with nonholonomic constraints cannot be stabilized by a static continuous-state feedback. In this paper, we consider a regulator problem of controlling a four-wheeled mobile vehicle from an arbitrary initial state to a target point, and apply the update type feedforward (FF) control input generation using the time-polynomial method (UFFT) to generate FF control input. The UFFT method is proposed as a control method for a crane system with restrictions on drive system, iteratively computes the control input at a constant time interval to reduce a residual sway. The generated FF control inputs are robust to modeling and quantization errors. The effectiveness of the UFFT method when applied to a four-wheeled mobile vehicle with front-wheel steering and rear-wheel drive is confirmed by the simulation.

In order to reduce the cost of an autonomous mobile robot, we are developing a system that can use cloud processing for the computation of autonomous movement and studying the risks involved in cloud processing. In this report, we evaluated the risk of the autonomous mobile robot during autonomous movement based on risk assessment, and examined the validity of the risk assessment based on the behavior of the program when communication with the cloud was interrupted or communication was delayed. In addition, we verified whether the autonomous movement program of the autonomous mobile robot in the simulation environment differs from that in the real environment. As a result, guidelines on communication risks were obtained

Ceiling is suitable location for robots to transport loads because Automated Guided Vehicle(AGV) and humans do not interfere. In a previous study, a ceiling mobile robot called HanGrawler 2 has been developed. It can travel at a high speed of 1.0 m/s to compete with ground vehicles. However, it sometimes fells at a high speed traveling. The purpose of this study is to improve the reliability of starting, accelerating, and traveling at high speed. Optical motion capture is used to observe crawler’s behavior of HanGrawler 2. The Observation of crawler’s behavior reveals that the crawler moves on an inflated trajectory at the high speed movement. In addition, the experiment results show that the collision is not caused by inflation, but by the push-in timing. Reliability of high-speed traveling was improved by installing an encoder and optimizing the push-in timing.

This paper focuses on cake topping performed by a soft gripper. In recent years, robots have been introduced into food factories for improving productivity and reducing labor costs. Robots are actively used in various manufacturing processes. Unfortunately, it is currently difficult to introduce robots into topping, which is a delicate task dealing with soft food materials. Therefore, this research challenges topping processes of cakes using robotic soft grippers. We fabricated fingers of soft grippers without gluing. Soft gripper fingers were cast in silicone rubber using gelatin as a sacrificial layer. A soft gripper was composed of a pair of soft fingers. A topping experiment was conducted using the fabricated soft grippers. First, we confirmed the usefulness of the soft grippers by conducting a topping experiment using strawberry samples. Next, topping experiments was conducted using real strawberries and cake bases. On the ’nappe’ base with a flat surface, strawberries could be topped straight in a uniform posture. Although the topping was successful on the ’piping’ base with uneven surface, it was found that the posture of the strawberries varied.

This paper proposes a novel method for controlling vibrational orbit of a flexible beam with a structural anisotropy. First, we introduce an analytical model consisting of a flexible beam whose flexural rigidity has anisotropy and a rotational actuator attached to the base of the beam. The rotational axis of the actuator is arranged to be nonparallel to either of two stiffness principal axes of the beam. Using the model, we theoretically derive the vibrational orbit of the tip of beam with respect to the synthetic wave input given to the actuator. Through the analysis, we show that Lissajous curve-based various orbits can be formed by tuning the control parameters, such as frequency ratio and phase difference. Finally, we demonstrate experiments. After confirming the validity of the proposed method, it is applied to a nonprehensile manipulation where three-degrees-of-freedom motion of an object is controlled by a single actuator.

Human joint is soft and flexible and has very important roles to play when human tries to walk, run or lift something. Many previous researchers have studied on the mechanics of human joint. In this paper, The dynamics of ankle joint can be expressed by a second-order linear differential equation constructed by inertial, viscosity and stiffness, these impedance parameters can be identified by system identification by testing the impulse response. the kinematic model of human ankle joint which takes flexible factors into account during walking and the method to identify the impedance parameters using snap-through buckling mechanism(snap motor), its design and available impulsive force are studied in this paper.

Removal of support materials is a major issue in 3D printing. In this paper, we propose a new method of molding rubber structures by using water or ice as support members or molds. After fabricating ice support members or mold in the freezer through Micro Block Assembly method, liquid rubber is poured in it and solidified in the freezer. Then, taking the rubber structure out of the freezer, water can be easily removed with no environmental impact. MBA is a new method of ice structure forming which we propose in this paper, where small ice elements are assembled to 3D structures. Two prototype examples are shown in this paper.

Soft robots need air supply tubes. However, if the supply tubes become long, Air pressure takes a long time to reach the end of the soft robot. therefore, operation frequency becomes slow down. In this paper, we propose a device that switches three flow passages self-excitedly by supplying continuous flow. This device helps to solve that problem by connecting the device to the end of the supply tube and switching flow passages just before the soft robot. The proposed device consists of two parts, one switches the effective flow passage by flat tubes and magnets, and the other shuts off other than the effective flow passage. The proposed device can be driven at low pressure because the switching motion is performed by a low-friction mechanism using magnets. The experimental result shows that the proposed device can change flow passages and switching frequency can be controlled by the flow rate.

In this study, we aim to establish a three-dimensional printing method for Inter-Crosslinking Network (ICN) structure gels in which magnetic powders of soft and hard magnetic particles are dispersed. Among various vat photopolymerization methods, such as the free-surface method, we investigated the modeling conditions and solution viscosity in a 3D printer employing the constrained-surface method and the optimal viscosity for modeling solutions. We investigate and report on the Mechanical properties of our 3D printed ICN magnetic gels.

3D printer plays in various fields, and study on strength and molding accuracy of 3D printed objects are being conducted. However there has been almost no study on changes in size and mass due to hygroscopicity. In this study, changes in size and mass of 3D printed objects that absorbed moisture have been measured. Based on these results, we have analyzed the tendency of size and mass changes and shown the tendency of size shrinkage to change depending on the amount of moisture in the modeled object. We have also confirmed the characteristics of size and mass changes depending on the internal structure differences.

The need for micro-object manipulation techniques is becoming important in the engineering and medical fields as objects become increasingly miniaturized. Laser manipulation has become an indispensable tool for research in various fields. On the other hand, laser manipulation requires high-power, high-quality infrared lasers, which have poor energy conversion efficiency. Therefore, we aim to develop a micro-object manipulation technique that can be driven by low light energy, using light-responsive swimming microorganisms that are expected to achieve highly efficient energy conversion. First, we conducted experiments using lasers to control the movement of Euglena. Next, microfluidic devices for object transport were fabricated. Finally, we integrated these devices and conducted experiments on object transport by Euglena, suggesting that object manipulation is possible at low light energies.

Drones are expected to play an active and important role in the logistics and infrastructure inspection fields. In non-GPS environments, drone have to obtain its position using self-position estimation technology. However, it is difficult to compare the estimation performance because test methods are not standardized. In response to this problem, we have started to propose standard test methods for self-localization of drone. In this study, a simulator for considering the test environment is developed by using Unity and the simulator can cooperates with the ROS (Robot Operating System). Three types of SLAM (Simultaneous Localization And Mapping) are compared in some environments prepared in the developed simulator.

When considering a mobile robot that can move and work on uneven terrain such as farmland, the following three specifications are required to enable both movement and work: the body must always remain horizontal, the height of the body can be changed, and the robot can move in all directions. We proposed a rough terrain mobile robot using a 4-axis differential module, a mechanism that can satisfy these requirements. Therefore, the specifications of the uneven terrain mobile robot, fabrication of the leg 1 module, and measurement of torque and force by attaching a 6-axis force sensor to it were conducted.

In this paper, we developed the control method for the fabric actuator based on force distribution information. The fabric actuator is one of the latest soft actuator, and it is believed to be used wide applications, but the reasonable control method has not been developed. We include the tactile distribution sensor to the system of controlling fabric actuator. It is cleared that controlling fabric actuator using force distribution information makes us to operate the fabric actuator more easily.

To train for vitreoretinal surgeries, such as peeling of the inner limiting membrane, a surgical simulator with an elaborate retinal model and contact detection method is necessary. Quantitative evaluation of forceps contacts to the retinal model is required because excessive forceps contact can cause loss of eyesight. This paper presents a detection method using image analysis to detect forceps contact with the retinal model. An evaluation system for contact detection of forceps was constructed, which includes an optical system incorporating an 850 nm LED, a CCD camera, and an infrared pass filter placed under a retinal model. This system observes the reflected light of infrared light at the forceps tips. The luminance value array of the forceps part in the images changed before and after the forceps contact. From the evaluation results, forceps contact was detected at a load of less than 0.1 mN within a radius of 2 mm.

This paper reports on the proposal and implementation of a method for fabricating hydrogel structures using a micro 3D printing system equipped with a precision leveling mechanism by thermal imprinting. Creating functional 3D microstructures such as sensors and actuators at the microscale is expected to contribute to the advancement of cell manipulation and cell culture systems. However, conventional 3D printing techniques have faced difficulties in controlling the gap between the building plate and the bottom of the substrate. A thermoplastic resin was used on the bottom surface of a 3D printer, and the surface shape of the building plate was transferred to the bottom surface to achieve a uniform distance between the two. With this method, the gradient of the height of the fabricated object within a 2mm square improved from 0.63°/mm in the x-direction and 1.03°/mm in the y-direction without thermal imprinting to 0.35°/mm in the x-direction and 0.31°/mm in the y-direction with thermal imprinting.

A robot system for automatically planting potato seed potatoes is being investigated. First, a survey of the method of planting potato seed potatoes was carried out. Subsequently, actual cultivation experiments were conducted to confirm the planting method. After that, based on the survey results and cultivation experiment results, a conceptual design of the robot systems was carried out. In this report, the results of investigation of the planting method, the results of cultivation experiments, and the results of the conceptual design of the robot systems are reported.

Recently, the SDGs, the Paris Agreement and carbon neutrality in Japan have been announced, and there is a worldwide demand for the reduction of CO2 emissions and the use of clean energy.

Therefore, we have focused on electric small mobility, which is easy to develop infrastructure and can be used as a means of transport using 'electricity' in depopulated areas. We have developed the 'pico-EV', which is driven by small amounts of electricity, and the 'electric-assist kickboard', which is an improved version of the electric kickboard with reduced electricity consumption. The amount of CO2 emissions is calculated from the amount of electricity consumed, and the CO2 emissions of electric small mobility vehicles when running are measured and the trends are discussed.

The. Previous work has been done to predict the lifetime of complex nonlinear systems, such as lithium-ion batteries, from early cycles. Therefore, I conducted research using deep learning to obtain more accurate results. However, we report that LSTM, which specializes in predicting time-series data, did not produce results.

We describe the development of a cord-like robot with redundant degrees of freedom using differential modules as a robot that can work not only in Satoyama, which is difficult to enter, but also in various environments such as caves and burrows of living things. A cord-like robot has high environmental adaptability and running performance, and in order to take advantage of their greatest characteristics, it is necessary to achieve high torque and robustness that can withstand physical damage in harsh environments. The specific work content of this robot includes "three-dimensional movement, monitoring, and measurement", and since it is assumed that the robot cannot be directly seen, it is operated using a remote control interface.

Infrastructure maintenance is a major problem. Asa’s Prior research has proposed a measurement planning method to measure damages with the resolution requested for damage detection. On the other hand, there is a limit to the work efficiency of a single UAV. Therefore, we aim to improve the efficiency of inspection work by having multiple UAV collect data in parallel. In particular, we will examine the allocation of work areas and collision avoidance required when the number of UAV is increased. The allocation will be done by clustering based on the measurement position, which affects the travel time of the UAV. Then, the measurement position corresponding to node of the travel route is changed to a position that can be kept at a distance between UAV and that can be measured with the required resolution.

A spherical camera inspection system was proposed for low-cost and efficient nondestructive inspection of tubular infrastructure. A transparent capsule-shaped inspection device with a built-in omni-directional camera was developed and rolled into a pipe. The inspection device rolls inside the pipe and captures images of the inner wall of the pipe as it moves. The obtained omni-directional images are image-processed to create a development map of the inner wall of the pipe. Verification experiments were conducted on the proposed system. The inside of a simulated pipe was photographed with a prototype inspection device, and a development map of the pipe inner wall surface was created. As a result, it was confirmed that the proposed system can produce appropriate development maps.

Social infrastructure is in urgent need of inspection and maintenance due to aging problems. In order to solve this problem, research and development of inspection robots is underway. The images captured by the camera mounted on the inspection robot include blurred images. Visual inspection of concrete measures crack widths of 0.05 mm or more, so if the image taken by the inspection robot contains blurring, the crack width cannot be measured accurately. Therefore, it is necessary to develop a system that automatically distinguishes between blurred and non-blurred images. In this paper, we describe the discrimination evaluation by quantitatively representing the blur of the image taken by the inspection robot. We evaluate the proposed method and reveal suitable features by discriminating blurred and unblurred images using a crack detection system.

When you get in automatic wheel chair, you may feel discomfort. This study assumes that the cause of the passenger’s feeling of anxiety is the surrounding environment, and aimed at generalizing passenger discomfort when the electric wheelchair is traveling autonomously．For this purpose, an experimental robot based on an electric wheelchair was used. By attaching 3D LiDAR as environmental observation devices to it, distance data of the robot’s surrounding environment is acquired. The acquired data is subjected to multiple regression analysis and regression analysis using a neural network to estimate the passenger’s sense of anxiety.