In passive walking, robustness against fluctuations in physical conditions is important. In previous studies, the method to simultaneously adjust joint stiffness and viscosity was reported for robust passive walking. However, this method was analyzed only through numerical simulations, and the mechanism to adjust joint stiffness and viscosity was not considered for an actual system. To develop the mechanism to simultaneously adjust joint stiffness and viscosity, this paper incorporates a muscle model that changes viscosity into both legs of a compass model and verifies its usefulness through simulation. In the simulation, varying the slope angle and joint mass, and the robustness of walking is compared with other systems.

This paper discusses methods of quadruped stability control, and introduce some new examples with inertial assist system, where high speed spinning wheel produces a gyro torque to keep whole robot stability. One example shows commercial gyro stabilizer is simply attached on the body and make a trot walking stability. Second one has an actuator to rotate stabilizer. The third uses simple wheel with axis tilting actuation. Later two can make a pace walk which needs more torque to stabilize. These stability control methods have an advantage that no landing point arrangement is needed, where sometimes arrangement can not be executed because of terrain conditions. Also, the body does not make an acceleration to produce a dynamic stability, which means the robot can go straight without any shaking motions.

Our previous research has focused on the development of quadruped wall-climbing robots that use suction cups. However, these robots have only been successful in climbing slopes to a limited extent due to their simplistic forward control strategies. This study aims to improve the walking strategies of quadruped wall-climbing robots by considering their body design and expanding their range of motion. Among different attachment methods, suction cups are preferred for their adaptability on smooth surfaces. However, the use of suction cups limits the degrees of freedom for the supporting legs during movement. To address this issue, this study suggests adding extra degrees of freedom to the robot’s body to compensate for this limitation, thereby increasing redundancy in the walking strategy. The first step involves simulating the changes in torso trajectory when introducing additional degrees of freedom to the robot’s spinal section.

In this paper, we considered the case of a bipedal robot carrying objects and people, and formulated the range of robot body acceleration at which the load does not roll and slip. In the robot that carries objects, it is essential to carry the load without falling it. To achieve this, it is important to control the rolling and slipping of the load. Based on a linear inverted pendulum mode (LIPM), we formulated a step length and center of gravity (CoG) height to prevent the load from rolling and slipping. In simulation, We confirmed that the load could be prevented from rolling and the slippage of the load can be reduced within the range of step length and CoG height we formulated.

The leg mechanism is a major factor in the performance of legged robots. Among various methods, leg mechanisms using a one-degree-of-freedom link mechanism have the advantage of easy control. However, design of the link mechanisms is demanding work because of inherent nonlinearity of the link mechanisms equation. Despite many methods are proposed in order to overcome this issue, it is difficult to apply these methods to design link leg mechanisms because existing methods use desired trajectory as objective, but trajectory is also design target in link leg mechanisms design. Thus, we propose a new design method that do not need desired trajectory as input and uses genetic algorithm. In this paper, we design link leg mechanisms through single objective optimization as preliminary experiment before design through multi objective optimization.

In this paper, we aim to reduce energy consumption during walking by applying the passive walking principle to biped robot that performs dynamic walking. First, a new equation of motion is obtained by treating the gravity component received by each link of a 4-link passive walking model with knees and arc feet as the output of a dynamic walking biped robot moving on flat floor. To account for the dynamic effect of the passive walking mechanism on the dynamic walking biped robot, the appropriate dimensions of the arc feet are derived from the equation of motions. In walking tests with the actual robot, the supporting leg tilted backward, resulting in a fall. However, the walking conditions are different between the passive walking model and the actuator biped robot, and it became clear that the coordinate system method and the output addition method need to be verified again.

This study provides a dynamic simulator aiming for stable operation of an autonomous distributed system. Multiple linked single-leg units were modeled in ROS/Gazebo, enabling the emergence of walking functionality by autonomously synchronizing leg movements. Adjusting contact parameters at the foot-ground contact point allowed for accurate simulation of real-world movements. This study constructed an environment to evaluate walking performance of an autonomous distributed system.

In this study, in order to improve the walking speed of a hexapod robot using the Follow-the-Contact-Point(FCP) gait control, the walking speed of the hexapod is examined by changing the speed of the swinging leg and the supporting leg. The maximum walking speed in this experiment is about 3.5 times faster than the walking speed of the previous study. The results also showed that the swinging leg speed is more significantly related to walking speed than the supporting leg speed.

Quadruped robot is effective for stable movement on uneven terrain such as disaster sites and construction sites. The leg mechanism of such a robot requires asymmetrical output characteristics, as it needs to exert a large downward force on the foot when supporting the body weight with a supporting leg, while only exerting a small upward force during swing leg motion. In this study, we developed a mechanism to assist the joint torque during supporting leg motion using a spring, which can be attached to the pitch axis drive motor shaft at the leg root where the weight of the main body is significant for movement on uneven terrain. We confirmed the effects during swing and support phases, as well as the spring effect. Additionally, using a quadruped walking model, we verified that the use of torsion springs improves coordinate accuracy during walking.

Improving mobility is necessary for the wheel-legged mobile robot to perform various tasks in different environments. This requires calculating the center of gravity, and one of the parameters of this calculation is the floor reaction force on the wheels. A method using frequency characteristics for determining the floor reaction force has been proposed. However, the conventional methods take a significant amount of time to obtain these frequency characteristics. Given that the load on the foot of robot changes, there is a need to rapidly obtain the frequency characteristics. Therefore, to address the issue of the conventional method, we propose a method for obtaining the frequency characteristics using the sliding discrete fourier transform (Sliding DFT) in a sequential manner. From the experimental evaluation, we confirm that the Sliding DFT can reduce the time for obtaining the frequency characteristics.

In recent years, legged rescue robots have been attracting attention. In disaster situations, the legs of the robot may fail, but the robot must still be able to perform its mission. However, few studies have been conducted on fault-tolerant gaits of quadruped robots using deep reinforcement learning. The objective of this study was to create a gait that can cope with failures during walking. The two cases of failure are considered, a fixed motor and a free motor. We used deep reinforcement learning to create a gait that allows the robot to continue walking without tipping over, even if it suddenly breaks down while walking. The results of the validation in the simulation showed the effectiveness of the proposed method.

This paper investigates the mathematical modeling and basic motion characteristics of a heptagonal passive rotating locomotion robot formed by connecting seven rigid frames via viscoelastic joints. The robot passively rotates on a gentle downhill while maintaining the body shape using the elastic forces around each joint, and has the ability to change the locomotion mode by reshaping its closure according to the road conditions. First, we describe the model assumptions and develop the mathematical equations of motion and collision. Second, we numerically show a typical passive rotating motion on a downhill and analyze the motion characteristics by changing the slope angle while checking the conditions necessary for stable motion generation.

The omni-roller is widely known as a means of omni-directional movement for mobile robots. In a related study, kinematics with the axial direction of the rollers fixed to the direction of the center of gravity of the machine body has been derived. This study introduces a proposed verification method for the general kinematics, which assumes that the position of the roller is arbitrary.

In order to restore the transportation network of goods and work vehicles blocked by landslides, it is effective to construct a temporary road by backhoe. Automation of the backhoe requires environmental measurements, path planning that involves modifying a rough terrain to reach the destination, and motion planning that performs the construction according to the plan. However it was unachieved to plan path and excavation to reach the destination with modifying terrain. Here we developed to simultaneously plan routes that satisfy the conditions and target modified terrain, allowing the selection of modification and movement policies according to the terrain to the destination by introducing cost function of distance, slope and the amount of excavation.

In Japan, the number of farmers is decreasing, and the population is aging by the declining birthrate and aging population. One of the solutions to cope with this problem are the autonomous weeding robots to reduce the burden of agricultural work, and many types of devices have been developed in various research institutes around the world. However, the actual weeding work fields are a harsh dynamic environment with steep slopes, so that it is difficult to make robots run autonomously, and it has not yet been put into practical use. In this study, we proposed a control system that suppresses skidding caused by driving on steep slopes to improve the straight-line stability of the robot.

This research proposes a new wheel design for mobile machines running under bumpy, fragile, and steep slopes such as rocky areas, icy fields, and bumpy off-roads. The wheel is composed of a minimal Tensegrity structure with three rods and an axle hub, which is a part of the Tensegrity structure and passes rotational force onto the wheel. Tensegrity has many unique features, such as deforming as a rubber ball while contacting the external surface with its rigid rod's endpoints, which are helpful as a wheel. The proposed wheel is expected to perform better than the other wheels, such as rubber tyres and crawlers, by including the axle hub as a part of the Tensegrity. In this report, we explain the details of the proposed Tensegrity wheel. We also show the results of the experiments by using a four-wheel drive mini motorcar with the proposed wheels. We compared the wheel with a "wheg" similar to the proposed wheel except for having no tension and its rods fixed. The results show the Tensegrity wheels' high performance in climbing gravelled hills and steps.

“ d-FlexCraw ”is a novel robot featuring a dual-body configuration with highly flexible tracks, allowing versatile three-dimensional curvature. The tracks adapt to terrain via a vertebral body with passive joints and multiple degrees of freedom, manipulated by wires. However, maintaining posture proves challenging due to external forces. This paper proposes a joint angle-based control algorithm to address this issue, aiming to enhance autonomous control and alleviate human workload. Our implementation focuses on evaluating straight-line performance, showcasing the potential of our approach.

In recent years, the frequency of natural disasters in Japan has increased, creating a need for construction robots to work in disaster areas. In the case of landslides, three important tasks are excavating, loading and transporting earth and sand. Traditionally, these operations have required several units of several types of heavy construction machines. However, it is difficult to access disaster sites. To solve this problem, it would be beneficial if a single machine could perform all three tasks. By adding digging and loading functions to the crawler, the work can be carried out in a compact body without the use of an arm. The “Antler” concept is to build a robot that uses track shoes for digging and loading. In this study, a robot equipped with the shoe proposed in the previous study was built and tested for digging, loading and walking performance. The feasibility of the robot was verified.

Planets such as the Moon and Mars have soft ground covered with sand. In such terrain, a rover that does not get stuck easily is required. Crawler-type rovers have good traveling performance on soft ground, and improving the traveling performance of crawlers-type rovers will enable exploration of previously unexpected area. One of the features that improves the traveling performance is the lugs. In this study, we focused on the inclination of the lugs. We consider that the inclination of the lugs increases running resistance and sand scraping force, which may improve running performance. This study examines effective conditions for crawler lugs and considers the development of a planetary exploration rover with excellent traveling performance in soft ground.

The lunar surface is covered with fine sand particles called regolith, which is soft ground. In this environment, the current lunar exploration rover is required to traverse slopes with an inclination of approximately 20 degrees. As a moving mechanism for the rover under such conditions, a crawler-type moving mechanism that is less likely to get stuck on soft ground can be considered. However, the disadvantage of crawlers is that their movement function is completely lost when the crawler belt is cut or detached. Therefore, in this study, we improve the structure of the crawler's drive wheels and consider the development of a mechanism that can move using only the drive wheels without significant deterioration in movement performance even when the crawler belt is cut or detached.

This study aims to develop a mobile robot to follow a specific human while avoiding obstacles and reaching to destinations. The behavior models for robot navigation are obtained by deep reinforcement learning in simulated environments representing simplified real-world conditions. The mobile robot in this study assumes to equip an RGB-D camera as an external sensor. The trained behavior model is then implemented in the actual robot for real-world navigation. Detecting and tracking of a specific target person is performed by combination of YOLO and SORT algorithms. Additionally, the RTAB-Map, which enables online simultaneous localization and mapping, allows the relative navigation from the current position without prior mapping.

This study presents the development of an autonomous control method of an angle of subcrawlers so that a vehicle body can maintain level when moving over a mountain or valley on a continuous ramp course, which is one of the test courses for disaster response robots. The developed controller assumes to have a terrain model beforehand and determines an angle of the subcrawlers by geometrical constraints. We verify the proposed controller with numerical simulations.

Crawler robots are mainly used as ground robots on uneven terrain such as disaster areas, but they lack redundancy and cannot move immediately if the crawler belt breaks. Therefore, the concept of modular robots has been introduced to crawler robots to reduce the risk of failure through a self-repair function. In addition, the load on all modules is equalized by the module placement optimization algorithm, preventing the recurrence of failures due to a large load only on a specific module. The module placement optimization algorithm was significantly superior to previous studies evaluated with a t-test at a significance level of 0.05.

In this paper, we describe the made of a crawler type of omni directional mobile robot whose belt uses a material with orthogonal anisotropic friction. Omnidirectional mobile mechanisms are used in robots that perform complex movements. However, In order to move in Omnidirectional movement, it is necessary to use multiple parts to eliminate friction in the direction perpendicular to the direction of rotation crawler belt, which increases cost of assembly. Therefore, by using materials with different coefficients of friction in the orthogonal directions for the crawler belt, the number of parts can be reduced. The robot consists of four crawlers whose belts are made of materials with different coefficients of friction in orthogonal directions to enable movement in all directions, a loading platform, and electrical circuits for control and status monitoring.

This paper describes the development of crawler robot for moving on rough ground. The pneumatic caisson method is a method used to construct underground structures such as foundations for piers and pumping stations. This method has a significant negative impact on the human body, so robots are being used to unman the work. However, robots currently in operation have the problem of getting stuck on uneven ground at the work site. To solve this problem, we developed a crawler robot that can move in multiple directions. A prototype was built, but lateral movement was not successful with any of the fabricated track shoe of any shape.

In a large-scale environment or during long-term use of SLAM, increase in data size and computational cost is crucial problems. Therefore, the need arises for maps with reduced data size, such as graph maps. Furthermore, it is essential to identify free space of the environment with low computational costs prior to constructing maps. Proximity point is a type of keypoints of point cloud that can be utilized for data reduction. In this paper, we propose a new method to extract free space with low computational cost based on a geometric property of proximity points. Subsequently, we construct 2D graph maps based on the modified Growing Neural Gas (GNG) algorithm which ensures that the edges do not interfere with the obstacles. Experiments are conducted using six sets of data acquired in an outdoor environment.

Due to the diverse shapes of robots tailored to their specific applications, conducting a uniform risk assessment is challenging. This research reevaluates the risk assessment for field and outdoor work robots, such as those used in agriculture, equipped with sensors to enable autonomous movement. It highlights the significant role that errors and losses in self-position estimation play in robot safety. Utilizing the initial risk assessment method in accordance with ISO 13482:2014, the study carried out a risk assessment for agricultural robots. By adding scenarios to anticipated damage locations, it became easier to anticipate risks and damages, offering clearer insights into risk identification. The findings indicate the need for further development in cloud robotics to emphasize the importance of a robot’s selfposition, aiming to create systems capable of handling more accurate self-positioning while maintaining possession of self-position.

Accurate vehicle attitude angle estimation is essential for automated driving and mapping. Therefore, we focus on the attitude angle estimation method using multiple GNSS antennas. In this study, a general-purpose IMU is used to interpolate the points where satellite signals cannot be received, with the aim of increasing the accuracy and reducing the cost of attitude angle estimation using multiple GNSS antennas. The integration with a general-purpose IMU is performed by using EKS to achieve higher accuracy. However, these methods assume that errors follow a Gaussian distribution, and errors in attitude angles calculated under the influence of multipath do not follow a Gaussian distribution, which causes errors to increase. By eliminating the attitude angles affected by multipath through reliability judgment, we aim to improve the accuracy of attitude angle estimation using EKS.

In recent years, research on automatic driving using LiDAR and 3D point cloud maps with high-precision information on the road periphery has been active. However, the construction of high-precision 3D point clouds is costly and affected by the environment. Therefore, in this study, we proposed a position/attitude estimation method that applies EKS and FGO to improve the accuracy of position/attitude estimation using general-purpose GNSS/IMU through post-processing. From the evaluation test, it was confirmed that the method can construct a 3D point cloud with the same accuracy as that of a point cloud construction GNSS/IMU using multiple GNSS antennas by applying EKS and FGO.

For autonomous navigation of so-called personal mobility robots, we have proposed a localization method using an existing cadastral map. As a result, a personal mobility robot was enabled to move toward a destination autonomously without building the environmental map beforehand through SLAM. However, obstacles in the environment are not included in the map. Therefore, the robot can not avoid obstacles ahead. For this problem, we allow the robot to reflect the obstacles observed by 3D LiDAR in the map through data association. Through the data association of the cadastral map and sensor information, the robot avoids the obstacles by repeatedly planning the path in the map. In the experiments, we show that the robot is able to move toward a goal destination autonomously by avoiding obstacles.

We conducted kinematic positioning experiments of Centimeter Level Augmentation Service (CLAS) positioning and RTK positioning, compared the location data of both, and verified the accuracy of CLAS using a slider that moves in a straight line at low speed. The measurement errors of CLAS positioning were statistically analyzed for the obtained positioning points. As a result of the verification, a bimodal peak distribution was obtained.

We are developing an autonomous mobile robot based on SLAM(Simultaneous Localization And Mapping) technology that can automatically explore environments where topographical information is unknown. In this study, we developed an algorithm to autonomously locate the highest radiation dose point without colliding with any obstacles, using radiation dose as a sensing target. The robot can autonomously avoid obstacles by using a path generation algorithm based on the Dynamic Window Approach. Radiation dose is detected by a GM counter, and the ROS (Robot Operating System) is used to link the robot with an autonomous mobile system to enable automatic acquisition of radiation dose data. The algorithm was verified by running the robot in an actual environment, and the robot was able to locate the point with the highest radiation level while avoiding obstacles in the unknown environment.

2D markers have been installed in the environment to indicate the reference points to robots. However, different robots use different markers. If the marker specification is standardized, the markers can be shared by the robots once they are installed. This paper reports the status of standardization of the “Shared Marker” as a building facility and exemplifies the self-localization methods.

Spare Keypoint Scene Coordinate Regressor (SKSCR) represents a learning-based approach to visual localization, aiming to overcome challenges posed by structural methods, including concerns related to storage, speed, and privacy. While SKSCR demonstrates competitive performance with structural methods across various benchmarks, its efficacy diminishes in scenarios with limited training data, inherent to deep learning models. In this study, we explore the integration of Neural Radiance Fields for synthesizing keypoint descriptors, providing a potential enhancement to SKSCR’s performance in situations where data availability is scarce.

In this paper, we improve a robust SLAM for dynamic environments by using object detection considering potential dynamic objects. Generally, dynamic environments have a negative impact on SLAM because correspondence between frames is wrong and constructed map includes ghosts. Conventional methods resolved this problem by predefining dynamic objects and removing them, however these cannot handle degenerate environments. Therefore, we define potential dynamic objects to increase the number of features as much as possible and to improve the problem of degeneracy that occurs in conventional methods. In this study, we use instance segmentation, which can detect objects more accurately at the pixel level, and verify its effectiveness.

The goal of this study is robust self-position estimation in environments where GNSS is not available; among the environments where GNSS is not available, this study targets an underground parking lot that is used by many people. In order to achieve highly accurate self-position estimation in underground parking lots, experiments were conducted using markers on cylindrical bases equipped with QR codes.

In this study, a smooth three-dimensional environmental map was constructed using the Kriging method that is one of the spatial interpolation estimation methods, and the gradient information of the map was used to conduct path planning using the A* algorithm considering hill-climbing. In the experiment of constructing a three-dimensional environmental map, it was possible to construct a qualitatively accurate environmental map on the floor surface, although it was estimated mountainously near the wall. In addition, path planning by the A* algorithm that was considering the cost of a robot’s hill-climbing was possible to conduct after extracting the movable area using the gradient information of the constructed three-dimensional environmental map.

In load-sensitive transmissions, two-way clutches are used to interrupt the transmission of force from the output shaft side to the low reduction ratio reducer. This paper proposes a two-way clutch with a co-axial double-screw mechanism, which enables to use in linear motion mechanisms and to reduce friction. The co-axial double-screw mechanism consists of the first and second screw, which are co-axially installed. To lock and unlock the screws, backdrivable sliding screws whose lead angle difference is 5 deg are used. In the experiments, we developed a prototype model of the double-screw mechanism, and verified locked and unlocked motion of the proposed mechanism using the weight of the slide screw itself.

A karakuri automatic guided vehicle (AGV) is proposed. The developed karakuri AGV has karakuri steering mechanism with a two-layer cam, rack-and-pinion and Ackermann-Jeantaud scheme. The karakuri steering mechanism without electronic control enables it to turn to the left and right. From the experimental results, it was found that the karakuri AGV can travel along the target route. The effect of the steering mechanism and payload on the driving route was also evaluated. The maximum average error for the target turning angle was 2.4 degrees. The maximum repeatability error was 2.7 degrees due to the load capacity and center of gravity position.

For humans and robots to coexist and cooperate, robots need to be inexpensive to lower the hurdle for introduction, and they need to make gentle contact with humans, objects, and the environment, so force control is a key technology. Torque sensor feedback is used to achieve contact, but the mounted torque sensor must be able to detect both small forces to realize skillful work while contacting an object and large forces such as collision. In this research, we focus on a method of using thrust displacement in the worm reduction mechanism when load torque is applied to the worm gear for sensing, and develop a low-cost, space-saving, and lightweight joint module in which the torque sensing function and drive unit are integrated without the need for additional sensor devices. Furthermore, two types of elastic support structures are used to provide two-stage characteristics to achieve fine to wide range torque sensing.

This paper describes a robot called “Tumbler-1” to measure the thickness of floor plates in oil tanks using an ultrasonic sensor. The robot has a structure in which the main body always stands upright by separating the center of buoyancy from the center of gravity. However, during the motion, the robot tilts due to fluid resistance, and thus, a simulator to estimate the posture in liquid was built as a part of the research and development. Finally, comparison between the simulated posture and the actual posture and thickness measurement experiments were conducted in an indoor pool.

This paper proposes achieving asymmetric force transmission by combining two tension-compression asymmetry beams with reverse V-shaped rods. Through simulations and experiments, we designed two types of beam beams: one that is soft in tension but stiff in compression and another that is stiff in tension but soft in compression. Experiments with the proposed V-shaped rods indicated asymmetric force transmission. Furthermore, we demonstrated the application of a vibration robot with the reverse V-shaped rods; we could observe both linear and rotational motion.

Bending mechanisms have been applied to microminimally invasive surgical robots because they are simple and advantageous for miniaturization. The authors have already developed a robotic surgical instrument that focuses on a 2-degreee-of-freedom (DOF) disk-stack bending mechanism for microminimally invasive surgery. We propose a 2-DOF microhand using disk-stack bending mechanisms. By applying this bending mechanism to a robot hand, an unprecedented, innovative microhand with a diameter of approximately 2–3 mm can be realized. In this study, a 5-times-enlarged model of the 2-DOF microhand with a diameter of 17 mm is prototyped, and basic motions of grasping and twisting are demonstrated by a manual drive. Furthermore, it is also possible to grasp parts of various shapes and hardnesses. This microhand is expected to be used as a soft hand.

Wire-driven systems generate tension by winding the wire, enabling strong forces up to the wire’s breaking strength to be transmitted over long distances. Furthermore, increasing the number of wires allows for an increase in the degrees of freedom of the controllable force. Robots incorporating the wire drive technology have been widely developed. In this study, we developed CubiX, a robot that utilizes the characteristics of wire-driven motion. CubiX connects itself to the environment via wires, which it uses to drive itself by winding them. Additionally, we conducted an experiment by integrating CubiX with other robots, thereby achieving object manipulation and spatial movement using wire-driven robots connected to the environment.

In tendon-driven robots, the ease of wire arrangement and its degrees of freedom are significant features. By ingeniously arranging wires, it is possible to construct a body that meets desired characteristics. Various wire arrangement optimization methods have been proposed; however, they often simplify the problem by assuming a constant moment arm or ignoring wire configurations with diverse relay points. In this study, we formulate a wire arrangement optimization problem with higher degrees of freedom, representing wires with starting points, multiple relay points, and end points. Utilizing multi-objective optimization that simultaneously considers feasible end-effector force and velocity spaces, we aim to achieve the desired robotic performance.

Modular robots, capable of forming various configurations by connecting multiple modules, are gaining attention as robots capable of adapting to diverse tasks and environments by altering their own configurations. However, individual modules have limited torque capabilities, making it challenging to lift several modules to form a serial-link manipulator to perform manipulation tasks. In this study, we propose the modular robot TYCOON (modular robot capable of Torque sYnthesized COllabOratioN), in which each module features torque transmission functionality allowing it to transmit its torque to other modules. This enables the collective exertion of greater torque than individual modules can achieve independently when multiple modules combine. We describe the complementary intermittent gear mechanism, which is essential for incorporating torque transmission functionality in addition to joint actuation functionality, which is commonly required for self-reconfiguring operations of modular robots. Furthermore, we introduce the configuration of a serial-link arm using torque transmission functionality and its application in manipulation tasks.

For tele-operated decommissioning robots, we proposed a retrofittable and disposable force sensor, and developed a linear-type force sensor. This sensor uses a straw and double clip, which converts the received load into water level displacement. This paper proposes adjusting the resolution methods for the force sensor. For the resolution adjustment, the following two factors are significant: (I)contact area between the straw and the double clip, (II)elasticity of the doble clip. Thus, we propose (1) the use of different double clip manufacturer (2) adjusting contact area using disposable chopsticks, and (3) adjusting elasticity by cutting the double clip’s back. Experimental results show that all methods can adjust the resolution.

We propose a fluid soft actuator that imitates the structure of the mimosa pudica. The mimosa pudica releases Ca2+ into cells when stimulated from the outside. The leaves can be closed by using Ca2+ as a trigger to change the amount of fluid inside the cell. In addition, the leaves can be opened again by restoring the amount of fluid over time. We propose the fluidic circuit that transmits stimulation signals, based on the signal transmission function of the mimosa pudica. Also, we propose the structure of the actuator that closes in response to a stimulation signal and returns to the open state over time. Finally, we examine the validity of the proposed method by experiments.

Some species of jellyfish exhibit intelligent behavior through phototaxis. This phenomenon is difficult to observe and analyze in an aquarium or the natural environment. To clarify this type of intelligence of jellyfish, the phototaxis should be visualized so that we can analyze it quantitatively. Therefore, we developed a mobile robot that is controlled by a living jellyfish. The small chamber with a jellyfish was mounted on the robot. the jellyfish's movement inside the chamber was recorded by a small camera. Optical flows, as bio-data, were estimated in real-time based on the recorded video data. As a result, we observed the robot can follow the light source by reflecting the bio-data.