Technology contests are based on the competition rules of a certain field. Participating in this is a kind of social implementation research. Authors developed a remote-controlled mobile robot to decontaminate the inside of a nuclear reactor building, and participated in the Decommissioning Robot Contest. In this paper, we report on robot research and development through technology-based contests from the perspective of social implementation-oriented research.

The goal of this research is to realize a robot development framework to prevent the abuse of robots using open source software. A basic framework has been proposed by demonstrating that program modifications can be identified by applying an image recognition method for malware and that a secure control system can be constructed using an FPGA and a microcontroller. In this paper, architecture of the secure processor, which will be the core of the control system, is considered, a prototype is made, and its basic performance is confirmed. Also, the functions and components required for this framework are discussed again based on the motion control experiments of a mobile robot.

While automation of production lines has progressed in recent years, it has not yet covered the handling of abnormal situations, for example, caused by the tipping of goods on a conveyor belt. In this research, a dual-arm cobot development environment is built and tried as part of construction of a mobile manipulation system using the dual-arm cobot with a freely movable wheeled mount. In this development environment, a system prototyping and trial environment based on ROS is provided, and in addition, robot behavior tools prepared by robot manufacturers are used together to enable effective simulation and actual operation of the robot. In fact, we present the planned and generated dual-arm motion of the cobot through the development environment, and confirmed that the motion of the 3D model on the computer could be displayed in conjunction with the monitoring of the total 14 DOF motion and its joint angles.

The high-definition (HD) map is widely used in autonomous driving to acquire correct road information during driving. At the same time, the HD map also needs to be updated promptly when road changes happen, such as establishing new traffic signs or the expansion of road lanes. Therefore, building a system that automatically detects road change areas is an essential issue need to be solved. In this paper, we had improved the accuracy of road sign detection then investigated change detection methods. We confirmed a proposed method to correct the position and orientation of the vehicle in real-time using road lanes detected from images with other data.

In order for autonomous personal mobility (PM) and pedestrians to coexist in the same pedestrian space, it is important for pedestrians to accept the PM when avoiding and passing each other. When passing each other, pedestrians decide their actions based on the behavior of the PM. Therefore, in this study, the avoidance behavior of pedestrians was investigated by experiments in cases against a pedestrian and a PM. The relation between pedestrians’ behavior and pedestrians’ acceptability was also investigated by the experiments. Results showed that avoidance-starting distance and passing interval affect pedestrians’ acceptability. Based on the results, requirements of a path planning method considering pedestrians’ acceptability was set.

Passenger comfort needs to be considered in the autonomous navigation of mobility scooters in pedestrian space. Previous works have shown that passenger comfort is affected by vehicle vibration and acceleration, and there are methods considering them. However, pedestrians are expected to exist continuously in large-scale facilities such as shopping malls, and it is not clear what factors may affect comfort in such space. Therefore, manual driving paths of mobility scooters were focused on in this study as paths comfortable for the passenger. The characteristics of the paths were investigated by analyzing paths collected with a driving simulator. Analysis results showed that scooter users tend to move in a free direction with no pedestrians, and paths with this characteristic was comfortable for the passengers. Based on the results, a design requirement of a path planning method for autonomous navigation in pedestrian space considering passenger comfort was set.

This study investigates an autonomous navigation method for personal mobility vehicles that is comfortable for passengers and can be used in crowded pedestrian spaces such as airports and shopping malls. Requirements for comfortable autonomous navigation were defined based on the characteristics of situation awareness related to comfort examined in our previous study, and a comfortable autonomous navigation method considering these requirements was proposed. The proposed method was evaluated by a numerical simulation and an experiment using a VR simulator to evaluate passenger comfort during autonomous navigation. The simulation and experiment results showed that the proposed method satisfies the requirements and can improve passenger comfort during autonomous navigation.

One of the causes of train delays is the reopening and closing of doors. Therefore, we propose an algorithm to determine when the doors can be safely closed within the time limit, based on the prediction of the passengers’ arrival time to the doors. The prediction of arrival time is based on the measurement of transit time. The closing door timing is determined by the list of possible closing times and the remaining time. In this paper, we describe the closed-door algorithm, experiments on subjects for predicting arrival time, and evaluation experiments of the proposed method. In experiments, the algorithm was shown to be effective for passengers who boarded the train at equal intervals.

As technology of in-vehicle systems progresses, electronization of automobiles also progresses and services provided by the automobiles are sofisticated. As the services are sofisticated, interaction between the automobile and human driver becomes important more and more. For this purpose, we have proposed advanced information input interface to the automobiles with a string-like vibration sensor on steering wheel to which drivers always touch in driving. In this study, we propose advanced information input interface equipped with differential property sensor instead of proportional property senser used in previous system, capable of discriminating grip forces for driving and input operation, via steering wheel while driving. Efficiency of proposed system is confirmed through information input experiments using previous system and proposed system for two in-vehicle devices, that is, a radio and an air conditioner as operational objects.

In order to predict highway travel time, we performed machine learning using vehicle detector data in a previous study. But the data of the vehicle detector is about 2 [km] apart, and the length of the traffic congestion cannot be accurately expressed. Therefore, it is expected to improve the accuracy by using ETC2.0 probe data which can obtain more detailed data. However, detailed data does not always improve the accuracy, and some problems are expected. In this paper, we will perform machine learning using ETC2.0 probe data, and verify its characteristics by comparing it with that of vehicle detector data.

Obstacle detection and avoidance performance tests were conducted to evaluate the performance of obstacle detection sensors mounted on drones in backlight and bad weather such as fog and rainfall. Through this, the performance evaluation items and test methods of obstacle detection sensors are reviewed to improve the drone's sensor and autonomous flight technology.

The remote driving system for the automobile is getting a great attention as the autonomous car widespread. The remote operation is regarded as the necessary function as a backup system in the case of malfunction of autonomous driving system. Generally speaking, it is impossible to ignore the transmission delay of the measurement signals, camera views and operation signals between the operated car and its operator. This delay induces the mismatch among the operation, the response in remote vehicle, and the operator’s view, and it makes the operation hard with degrading the accuracy and appropriateness of the driving. This study compares the driving performance of the two remote operation approaches to each other. The one provides the ’predicted view’ of the remote car to the operator. The predicted position considering the delay time is computed from the measured position of the remote car, the history of the operation, and vehicle model to compensate the transmission delay. The predicted view is generated from given virtual 3D environment and the predicted position. The other is a leader-follower teleoperation using a virtual car controlled by the operator locally and followed by the remote car with the path following control. The performances of these remote driving method are compared with each other in a driving simulator experiment.

In this study, we have designed and developed a coanda drone that combines the characteristics of fixed-wing drone and rotary-wing drone. The coanda drone is composed of a main body and four propulsion units. The propulsion tubes of the propulsors were developed based on the structure of a jet engine. A ducted fan is also installed inside of the propulsion tube to obtain high thrust. The drone’s dimensions are 191×121.6×121.6[mm], and it has a maximum total thrust of 5428[g] and a total mass of 1166.7[g]. The effectiveness of the mechanism of the drone was confirmed through flight experiments.

In this paper, we propose a method for transporting an object in a non-fixed state between the object and the aircraft or between the aircraft and each other using multiple caged unmanned aerial vehicles without a grasping mechanism. We also give an overview of the experiments using this method and their results.

The Rapid Application Development (RAD) concept is an approach for developing high-quality computer systems in a short period using Computer-Aided Software Engineering (CASE) tools. This paper aims to build a RAD framework for UAV development to enable flexible and rapid verification for various hardware and software configurations. This paper presents the first step of developing an auto-tuning system for a quad tilt-rotor UAV to rapidly complete the tuning of control parameters, which takes up most of the development time. In this paper, the RAD framework for UAV development is realized using the MATLAB/Simulink Embedded coder as much as possible, and a seamless development process between the software and the hardware implementation is achieved.

The wings of a flapping flying robot occupy a large amount of space and are easily damaged when the robot moves to the observation point. One method to solve this problem is to implement foldable wings. By using this method, the space occupied by the wing can be reduced by folding the wing into a smaller size. As a result, the wings are prevented from being damaged and can be spread out to fly away at the observation point. In this paper, we will reduce the weight of folding wings and fabricate two types of wings with different frame structures. The wings are mounted on the robot's body, and flight experiments and motion analysis are conducted. The results of the motion analysis suggest that the flapping frequency changes depending on the size of the wing deflection, and that the difference in the frame structure affects the wing deflection when the robot flaps.

We are developing a small flapping-wing robot to be used as an observation system in hazardous environments. In this paper, in order to investigate the effect of the wing shape of the butterfly robot under development on the lift, we analyzed the lift generated when flapping four models with different wing shapes. At the same time, the relationship between lift and flapping angle was evaluated by measuring the flapping angle during flapping. As a result, it was found that the shape of the forewing fore-edge and the diameter of the fore-edge frame affected the lift. In addition, the relationship was found that the lift in the negative direction became maximum after the flapping angle reached the upper dead center, and the lift in the positive direction became maximum after the lower dead center.

Multi-rotor UAVs are important for various applications, but they are driven by electric motors and have the limitation of short cruising time. In this study, the composition and control system of a UAV that achieves attitude control by vertical fins are proposed. The suggested UAV has a single rotor, which, using an actuator, can be converted into an internal combustion engine. Herein, the results of the dynamical simulations and tests are described.

The purpose of this research is to develop mechanical gimbal control system facing directional antenna at ground station and mobile station on drone each other sending 4K video in real time on several drones. I developed hardware and software for gimbal to make the antenna on the mobile station facing on the ground station. And I newly developed three points in order to Gimbal system of ground station incorporating drone to downsize, introduction of ground altitude, and support for drone turning using azimuth, and running several operational exams.

The recent development of UAV technology has popularized UAV-based systems. However, to use a UAV for package transportation, it is required to select a UAV that satisfies the required conditions of the transported package. In this paper, we propose a cooperative transport system using multiple Coaxial-UAVs.Especially, the aim is to achieve cooperative transportation in cases of unknown positional relationships. For this, it is important to know which axis of rotation the UAV acts on. This paper presents a method to estimate the positions of the UAV relative to the plant using a Neural Network. As a result, it was suggested that the estimation is possible under specific conditions.

Seabirds effectively use winds to gain airspeed and generate large lift force. It also saves energy and extends flight distance by gliding flight. In this research, aiming at the realization of a flying robot that enables long-distance and long-time flight, we measured the air volume in multiple directions using multiple small flow sensors and tried to estimate the surrounding wind speed and direction from the data.

Recently, multi-copters have been used in various areas, including outside and inside buildings. In the inside flight, airflow due to a wall or ceiling may affect multi-copters. This study aims to clarify the relationship between the airflow around the propeller and the force applied to the propeller near the wall using CFD (Computational Fluid Dynamics) analysis. This paper, first, briefly introduces the results of PIV (Particle Image Velocimetry) analysis and force measured in the experiment with an actual propeller. Then, this paper demonstrates qualitative characteristics of the relationship between the airflow around the propeller and the force applied to the propeller by comparing the results of CFD analysis with those of actual experiments.

In a previous study, mid-air retrieval with fixed-wing UAVs to a low-speed falling object was considered, and the approach trajectory generation using USFC (updating final-state control) was proposed. However, the effect of the delay on the control law has not been investigated for the real system. In particular, unlike the feedback attitude control by PID control using IMU, the trajectory generation using GNSS system is expected to have relatively large delay due to sensing and control calculation. In this study, a control law that allows for the delay caused by those processes was designed. Moreover, the effectiveness of the designed control law was confirmed under some conditions by simulation. In addition, based on the simulation results, an experimental system capable of parallel processing was constructed to realize the control by UFSC.

Multicopter is capable of vertical takeoff and landing, and it does not require a long runway. Therefore, in the event of an emergency such as a disaster, it is possible to quickly respond to grasping the current situation by aerial photography. However, ordinary multicopters are designed for takeoff and landing in a flat surface location. So it is difficult to take off and land on rough terrain such as a place where rubble spreads or a mountain slope, and there is a possibility of falling. In order to solve this problem, a mechanism that can land at any landing point is required. Therefore, we propose a landing gear using a linear actuator that can land on rough terrain such as steps and slopes.

In this paper, we develop a dynamic lead-lag mechanism to control the rotation of a butterfly-style flapping robot. The current butterfly flapping robot can only fly in a straight ahead. In this paper, we propose a dynamic lead-lag mechanism to control a butterfly flapping robot. In this study, we develop a lead-lag mechanism using a shape-memory alloy actuator, and verify the turning performance of the lead-lag mechanism by performing lead-lag motions while the glider equipped with the mechanism is gliding. As a result of the glide experiment, it was found that the glider turned with a yaw angle change when the lead-lag motion was applied during the glide.

GNSS is widely utilized for positioning of UAVs. As the UAV executes missions in the vicinity of facilities, the robust GNSS-based positioning is a challenging problem. In the previous research, a cooperative positioning system consists of multiple UAVs was introduced to support the positioning of the UAV works near the facilities. However, multipath propagation of GNSS signal caused by the facilities will deteriorate the positioning quality of system. In this research, a real-time path planning method that dynamically plans the paths of the UAVs of cooperative positioning system is proposed, to lead the UAV to leave the area where multipath propagation is detected for a more accurate positioning, while the simulation of proposed method is conducted. From the results of simulation the effectiveness of proposed method is confirmed.

In recent years, there is problem which increasing the number of space debris and risk of collisions with space debris in Earth’s orbit. In particular, due to low air loss and gravitational field, problem which collision with decommissioned spacecraft in geostationary orbit. Therefore, it is necessary to retrieve the decommissioned spacecraft from geostationary orbit. In process of recovery, it is vaital technique to estimate relative attitude motion of the target spacecraft. This study aimes to propose a method to identify the target’s attitude motion based on movies taken from other spacecraft at real-time. In this paper, we propose a method using Extended Kalman Filter and construct a system estimating the attitude motion using quaternion. Assuming that the target’s movement is nutation motion, we report the result of numerical examples and outline of ground experiment conditions.

In order to verify the behavior of the climber by the hybrid drive method, the climber with opposite and cross rollers with the same geometric and drive mechanism was manufactured, and their operating characteristics was compared. It was estimated that the action of the pressing friction force and the transmission friction force differs depending on the roller method, the starting current was higher in the cross roller type, and the contribution of the transmission friction force was large. The Space Elevator Challenge 2021 was participated in to confirm the climber's performance. In this report, the results of climber operation experiments and its challenge state were explained.

This paper presents a small mobile robot using uniaxial wave propagation mechanism. The wave propagation mechanism is driven by a single helical shaft driven with a single rotary motor. The prototyped robot is equipped with the dual wave propagation mechanisms to achieve steering locomotion on a plane. The developed robot was experimented to evaluate the mobility performance of the developed robot, and the experiments confirmed that the developed robot can demonstrate translational locomotion, steering, and spin turn on flat surfaces.

This paper presents the gripping characteristics of a passive spine gripper for multi-limbed free-climbing robots in extreme terrain. Although the gripper is mechanically-simple and efficient, its gripping mechanism needs to be clarified towards an optimal design. The characteristics were experimentally investigated by changing design parameters: finger number. finger length, and spine attachment angle on the finger. In the experiments, we evaluated the maximum pulling force when the gripper holds the irregular object. The comparative results showed that a larger number of the finger and long fingers provide better performance in the near-vertical pulling conditions. In addition, the result confirmed that the finger located on the opposite side of the pulling force largely affects the performance in the near-horizontal pulling conditions.

Currently, while a wide variety of experiments are being conducted at the International Space Station (ISS), it is also necessary to constantly perform the daily routines to maintain ISS such as cleaning and transporting supplies. In order to reduce the burden on astronauts and maximize their availability for more advanced missions, some mobile robots in ISS called Intra-Vehicular Robots (IVR) have been proposed to assist their tasks. As more versatile IVR, We propose to utilize the legged robotic technology that is capable of holding and walking by grasping the handrails for ISS crews. This paper particularly focuses on the development and performance evaluation of the gripping end-effector for that system.

Intra-Vehicular mobile robots are expected to assist astronauts in their work on the International Space Station (ISS) or future lunar gateway. Since such robots are required to carry cargo, open and close hatches, they need a articulated mechanisms for manipulating them and fixing the robot itself. In our laboratory, we have been developing a four-limbed climbing robot for moving while grasping the terrain surface with grippers mounted on the leg tips. In this paper, based on this legged climbing robot, we examine a leg configuration and gait planning method for a legged robot that can walk along the handrail in the ISS, and verify the proposed method by simulation.

On the International Space Station (ISS), astronauts are working on a variety of tasks. The space-saving environment of the ISS means that only a limited astronauts can stay on board. Therefore, the time available for astronauts’s activities is one of the most valuable resources. In the future, it will become even more important for future lunar orbit base and other human space bases.

JAXA has analyzed routine and repetitive astronauts’ tasks in the Japan Experimental Module (JEM) and are working to automate and autonomize them by IVR (Intra-Vehicular Robotics) so that the astronauts can focus on more advanced tasks. One of the routine tasks is the video shooting tasks. Int-Ball project was established with the goal of reducing the crew time for shooting tasks to zero using drone-type free-flying robot. And JEM Internal Ball Camera (Int-Ball) was launched in 2017 as first model. JAXA is now developing a second robot based on the issues learned from that launch. The second model will fly more autonomously take pictures and video shooting without crew time on various missions.

This paper elaborates on the experimental investigation on the traveling performance of a rigid wheel with a grouser attachment angle in loose soil. The experiments were carried out using the single wheel testbed, and we also conducted the PIV (particle image velocimetry) analysis to understand the dynamic soil flow field beneath the traveling wheel. The results showed no significant difference in the traveling performance, such as wheel slip, was observed for the change of the grouser attachment angle. On the other hand, from the results, the wheel slip, wheel sinkage, and slip line thickness are influenced by one another. The experimental analyses confirmed that the lower slip and sinkage are caused by the lower slip line thickness.

Small rover navigation needs to realise the direction of a landmark such as a mother lander. However, due to the hardware limitations, small rovers are hard to implement high accuracy perception algorithms. This paper proposeds a novel detection algorithm using saliency map. We focus on the peculiarity of an artifical object in natural environment, and study the estimation of the lander by analyzing the saliency map, painting, and linearity of the feature area from the image. We can detection landa by low accuracy perception algorithms.

In recent years, the number of space debris in the earth orbit has been increasing, and it is expected to remove the space debris with a small removal satellite. When capturing space debris with a robotic arm equipped with an actuator with a high deceleration ratio, the arm may be damaged due to the forced displacement caused by the inertia of the large mass due to poor back-drivability. In this study, we propose to incorporate a torsional elastic mechanism in each joint in addition to impedance control to buffer high-frequency loads that cannot be handled by impedance control. A prototype of the elastic mechanism was designed and simulated.

When capturing space debris, which is a non-cooperative object, there is not always a place where it can be grasped. Some capture mechanisms that wrap around the entire target are proposed. However, these mechanisms have a problem of failing in grasping due to the difficulty in predicting the motion in advance. The authors focused on "thigmotropism," a movement in the direction of growth that occurs as a mechanosensory response to a touch stimulus such as tendrils. This property can be used to wrap and capture debris adaptively. The authors have proposed a multi-link arm mechanism that consists of modules that operate by detecting contact and designed an operation algorithm to realize a thigmotropism-like adaptive capture. In this paper, a concept of this device and a design and prototype development of the mechanism are discussed.

The number of space debris in orbit is increasing. In order to suppress the increase of debris, Active Debris Removal (ADR) is necessary, and it is necessary to develop a capture technology for non-cooperative targets. A debris capture mechanism using a convex spring has been proposed, but improvement of the motion control of the mechanism is an issue. The author proposes attaching weights to the convex springs to improve the capturing motion. First, the bending propagation characteristics of a convex spring with a weight attached are clarified, and a mechanical model is derived. Next, the consistency of the mechanical model was confirmed by conducting experiments to measure the motion of a convex spring. Furthermore, the effectiveness of the weighted convex spring in debris capture was confirmed by conducting capture experiments on a fixed target. Finally, a control method of the weighted convex and a further improvement of the function by attaching the exoskeleton to the convex are proposed.

Hazard detection, simultaneous localization and mapping (SLAM), and terrain classification are crucial for autonomous navigation of the Moon or remote planets. In this paper, we integrate a 360-degree laser sensor (LIDAR) and an omni-directional camera in the co-axial arrangement to achieve the above three goals. The point-cloud information from the LIDAR is used to detect hazardous objects, create a high-precision map of the surrounding environment, and then localize the rover. The visual images from the omni-directional camera are used to add semantic information of the detected objects through the mask recurrent convolutional neural network (Mask R-CNN) technique. The performance of the integrated system was evaluated in a simulated planetary field.

Music is a tool that enriches our lives. Recorded electronic sound sources can be listened to on a daily basis due to the development of CDs and portable music players. However, listening to live music requires a performer, and there are few opportunities to listen to live music compared to electronic sound sources. As a method of listening to live music unattended, automatic performance devices such as music boxes have existed in the past. However, most automatic performances are often limited to playing melodies. Percussion instrument automatic performance devices require high-speed movement when playing drums, and it is difficult to create the speed at which the drums are struck. Therefore, percussion instrument automatic performance devices are limited to those that are relatively easy to produce sound, such as keyboard percussion instruments. Has been done. Therefore, in this research, instead of using a large drum as in the past, we used a pipe drum that uses a vinyl chloride pipe that has a relatively small diameter and is easy to produce sound, and whose pitch can be changed by the length of the pipe. Develop an automatic performance device for Tubulum.

We are studying human and robot ensembles as a highly collaborative work. In this report, we considered the tracking method and the detection method of the performance start point (Einsatz), and obtained the following conclusions. By comparing tracking by cascade, KCF, MIL, and MediaPipe, it was shown that skeleton detection by MediaPipe is desirable comprehensively from the frame rate, tracking stability, and the large number of feature points. In Einsatz detection using MediaPipe and MLP classifier, the optimum parameters were extracted by performing a grid search.

We are developing an anthropomorphic robot that can play the violin to enhance the communication ability with humans through its expressive performance. The bow force was not controlled in our previous report, which will cause alternation of the bow force and reduce the sound quality. This report introduces a robot’s right hand and its bow force control system. Each four fingers except the thumb have a chamber covered with a silicone membrane, which inflates when the air pressure inside the chamber increase. Air pressure of each finger can be controlled independently. By altering the air pressure in the index finger during the down bow, we can control the bow force. Experimental result shows that the system can reduce the fluctuation of the bow force from the target.

This study reports design and fabrication of the left hand of a violin-playing robot. Our previous left hand uses a link mechanism and a rack mechanism, which are driven by RC servomotors mounted on the left hand. Since this increases the mass of the hand, it was necessary to reduce the weight of the hand to allow the left arm to move fast. Each joint of the new left finger is driven by a wire, which allows us to install the actuator outside of the hand. The mass of the fabricated left finger is small by 30 % compared with our previous hand’s finger. We conducted a step response experiment and confirmed that the finger reached to the reference angle within 0.8 ms.

This study aims at building a system capable of automatically determining the bowing and fingering parameters for an anthropomorphic violinist robot. We adopt the reinforcement leaning for determining the bow speed and the bowing-direction under the constant bow-force condition, in which a neural network is embedded as a value function. In the previous report, the learning success rate was small, which might be caused by the reward values. Thus, the reword values were changed in this report. In addition, we applied the simulation method to a short musical piece of four bars. As a result, we found that more stable leaning results were obtained by using the new rewards values, and our proposed method can be applied to the four-bar musical piece.

The purpose of this study is to develop a Kendo-Kata robot to support elderly people who enjoy Kendo. In developing a kendo robot, it is necessary for Kendo-Kata robot to synchronize the movements with the user and to imitate Kendo-Kata movements. In this paper, we suggest a method to estimate the transition timing of Kendo-Kata movements, which is necessary for synchronization of movements.

In this paper, I describe an automatic musical instrument that uses cams to enable analog-style performance. The purpose of this paper is to propose a new musical expression and performance experience by creating a playback device that allows the user to be involved in the performance without any special skills. The contents of the research and experiments include the investigation of the instrument to be used, the basic experiment of the recording method using the cam mechanism, and other important mechanisms. Through this instrument, we will propose a new musical experience by enabling automatic performance that cannot be expressed by existing music boxes.

As one of the increasing eye diseases, glaucoma may cause vision loss, and surgery is performed to reduce intraocular pressure for its treatment. This research introduces a fabrication method for an artificial eye model mimicking the glaucoma eye structure comprised of the porous trabecular meshwork and the hollow structure of Schlemm's canal (SC). Furthermore, we integrated the eye model with a system electrically detecting the contact between a surgical instrument and electrode film formed on the bottom of SC, and confirmed the validity of its glaucoma surgery simulator. This simulator can be used for the surgical training of ophthalmologists and evaluations of new MIGS instruments.

In minimally invasive surgery, force sensing for surgical robots is highly desired. To solve this problem, we have proposed a force sensing method using string vibration for wire-driven forceps. The purpose of this study is to apply this method to a 5 mm forceps with a flexible joint. A flexible joint forceps is expected to have smaller friction. However, its maximum wire tension is limited due to its mechanical strength. To extend the measurable range of the wire tension, wire tension amplification mechanism is proposed. Force sensing experiments were conducted using the developed forceps. The results confirmed the effectiveness of the contact force sensing in two axial directions.

Ultrasonography is less invasive and safer than MRI or CT. However, image acquisition is dependent on the skill of the probe operator, and it is difficult for an untrained examiner to understand the three-dimensional structure of the organ. For the above reason, we use a robot to acquire ultrasound images and estimate the position of the probe using Visual SLAM. The obtained images are segmented and combined with position information to construct a three-dimensional model of the organ. In this study, the right kidney of the phantom was used as the target, and the results of position estimation and segmentation accuracy were high. The model of the right kidney was constructed based on these results.

Echography has advantages of small size, inexpensive, and low invasiveness. However, it requires proficiency in probe scanning to obtain desired echogram. Therefore, training equipment and simulated human bodies for echography education have been released, but there have been few case databases of patients. Therefore, in this study, we developed an ultrasound training system, ECOT. In this paper, we constructed a coordinate transformation system that correctly measures probe posture and a matching system that adds time information to the position and posture information to improve the matching accuracy of ECOT. As a result, we succeeded in obtaining the position and posture of the probe tip independent of the type and position of the measurement system. Even when the database was constructed from organs whose shape changes over time, we were able to construct a system that displays echographic images in the same way as the real body by adding time information.