A cooperative transportation with some position controlled mobile manipulators causes some problems. It is an internal force occurring in a conveyed object due to a position error. So we suggest a system using a wire to solve the problem. A carrier is fixed by using six wires and can build the system which internal force does not produce. In addition, we devised the revision method of a position error and the posture error from the transportation target position by the position error of the mobile manipulators that occurred when transported a carrier. And we inspected the method of this revision that devised.

In this study, we propose a control of a towed mobile robot equipped with servo brakes and no motors. The towed mobile robot can move forward by utilizing an external pulling force, and steers itself by controlling the servo brakes attached to each wheel. The control of passive type robot is more complicated compared to active robots which can move forward by simply applying voltage on actuators, so we propose a motion control method of the towed mobile robot based on the concept of feasible braking control region. Also, it is pulled from higher position by tether, so that tether does not touch and trail on the ground.

In multiple robots environment, it is an essential technique that the robots avoid collisions with each other. To achieve efficient collision avoidance, it is desirable that each robot avoids surrounding robots considering not only their positions but also their behaviors. We have developed spatially seamless local communication system in order to exchange behaviors between the robots. This system ensures spatially seamless communicable area towards robot's surroundings by rotating a transmitter and a receiver. However, a bit rate of communication using the system is insufficient. For this problem, frame structure with multiple carrier waves for transmission pulse has been proposed and improves the bit rate in the system by 10 times. In this paper, improvement methods of the communication quality for such communication system is proposed.

In modular robotics, the robot construction method by combining robot modules has been proposed. Although the modular robot system is expected to have great versatility in its shape, combining a large number of modules results in an increase of the total weight and torque deficiency. In this study, we propose a gravity compensation modular robot in order to solve this problem. This method gets rid of the limitation on number of modules that configure a robot, and enables the modular robot system to construct more diverse shapes. In this paper, we describe the method of gravity compensation, the control method (under consideration) of the module, fabrication of a prototype of the module and mention future works at the end.

In the present paper, the authors propose a reconfigurable multi robot system which consists of many robotic dolly units capable of transporting a heavy object in a group and a central unit for measuring the object's motion and distributing the acquired information to all dolly units. The system has no hierarchical control structure or supervising host computer, but is capable of achieving cooperative transportation by individual autonomous motion control of each dolly unit. A systematic calibration method of the position of each dolly unit and alignment of its local coordinate against the central unit's coordinate, and subsequent motion control algorithm to trace a target path given by a user are proposed. Result of the proposed methods is demonstrated by a series of experiments. Additionally, a feedback control scheme to reduce the effect of position estimation error is also proposed and demonstrated.

In this study, utilizing multiple mobile robots for a mega-solar power plant is proposed. The system is consists with two mobile robots and a parallel wire mechanism. Two mobile robots needs to align each other in coordination during inspection. In this paper, a method to estimate position each other using wire used for fault diagnosis.

This paper shows experimental results for application to step field of the robot that equipped with passive crawlers for stair climbing. A step field is designed and proposed by NIST to evaluate mobility of robots. The prototype robot “SWART-IV” that can climb stairs of 40[mm] in height and 45 degrees in inclination was tested on a step field made by toy blocks. Experimental results show that the robot can traverse the step field in some appropriate condition, but it is need to be controlled with visual feedback by a human controller.

In this paper, we propose a motion planning method for step climbing of a boarding-type mobility robot. We analyzed the mechanism of the boarding-type mobility robot. The climbing motion was split into three phases. We defined evaluation function as a linear combination of the cost functions. The next state of the robot was a state in which the value of the evaluation function is minimum. In this method, we used cost function to suppress seat tilt angle and power consumption in addition to mechanical limitations. We confirmed that boarding-type mobility robot was able to climb a step and seat tilt and power consumption were suppressed simultaneously.

This paper describes development of mobile robot using semicircular wheels on uneven terrain. It performs crawl gait on flat ground and butterfly gait to climb over debris. During crawl gait rotates the phases of the respective wheels shifted. During butterfly gait rotates all wheels in coordinate phase. Perform an experiment on the crawl gait and butterfly gait. We confirm respective gait method from the motion experiment.

This paper introduces the application of force control to the planar motion of the manipulator of a wheeled inverted pendulum assistant robot. The aim is to be able to control both position and force at the manipulator tip during execution of tasks like wiping tables or sweeping floors. To gain insight on the mechanism involved in planar manipulation, basic modeling using static equilibrium and geometric constraints is performed. An actual test apparatus is developed and tested to verify the validity of the model. Discussions and results are presented.

Robots using non-circular wheels have been studied as one type of robot that aims to search for rough terrain. Compared to circular wheels, noncircular wheels can be hooked on steps and irregularities, and a large thrust is obtained in the direction of travel. However, the robot using non-circular wheels has a problem that the car body moves up and down while traveling on a flat road. The purpose of this research is to develop a robot that can run on irregular ground and can travel without moving the body up and down during running on flat ground using non-circular wheel. In this paper, we report prototype design and operation experiment of rough terrain mobile using non-circular wheel.

We are studying an inverted pendulum type robotic wheelchair with the ability of climbing stairs. The structure of the wheelchair consists of a seat slider and two rotary links between the front and rear wheels on either side. As the control, we used a linear-quadratic regulator consisting of a center of gravity control mode and a rotary link control mode. This work is focused on method for restraining movement of wheels when it climbs stairs, which causes fall of wheel chair. We grasp the mechanism the movement is caused by mutual interference behavior between wheel and rotary links in climbing stairs. We proposed two methods for decreasing interference behavior by changing the target track of rotary links and confirmed the effect of restraining movement of wheels.

In this paper, we describe a development of a robot that plays the game “Red Light, Green Light” We have developed a detection method of the looking-back behavior of a human using OpenPose. Using OpenPose, we can calculate the distance between both shoulders. Then we can detect looking-back behavior of a human by comparing the distance with the threshold. We conducted experiments to evaluate the method. As a result, we could detect the behavior by using the nose point feature together with the distance between both shoulders.

In this paper, we describe development of a path planning method of the mobile robot to avoid a pedestrian approaching from front and pass by the pedestrian. When the robot detects a pedestrian, it calculates the trajectory to avoid the pedestrian considering the personal space. Then the robot passes by the pedestrian. Finally, the robot returns to the original trajectory. We conducted an experiment to evaluate the method with the three routes and two speed conditions of the pedestrian. As a result, the robot system functioned without problems.

The aim of this study is to recognize human emotional state by multiple physiological signals. Our final target is to develop a real-time emotion recognition system, it has multiple uses in our daily life. With accurate and objective emotion analysis results, our method can provide useful treatment effect information to use robots or some assistant apparatus which service activities of daily living (ADL). In this research, we used normative affective stimuluses to elicit subjects’ emotional change and collected multiple physiological data, then analyze data use the method we proposed to grade the emotional state. The innovation of our research methods: we construct a deep neural network to deal with the physiological signal to recognize human emotional states. The experimental results show that the proposed methodology provides a recognition rate of 80% for different emotional states. The calculated results reveal that significant improvements with other traditional methods, e.g. SVM, K-Means, and Naive Bayes.

In the previous research, a pressure presentation device was developed in order to reproduce the traditional instruction methods of calligraphy handwritings. By this device, instructions were carried out to the learner only when instructions were necessary, that is, if the brush overruns the reference handwriting, the pressure was presented to the student's wrist that holds the brush (Single presentation). In the experiments, the developed device was applied for calligraphy self-trainings, and instructions were carried out correctly without losing the learner's activities. In this paper, a new instruction method is proposed in order to decreases further the overrun of the handwritings. In this method, the pressure is also presented even before the brush tip reaches the end point of the reference handwriting. This is called Double presentation method. As a result, it was confirmed that Double presentation decreases overrun rather than Single presentation.

Computer mice and keyboards are most widely used to work with personal computers as pointing and typing devices respectively. As alternatives to a computer mouse, there are other types of pointing devices that use depth sensor or camera for pointing detection. However, current implementations are uncomfortable for usage because a user must raise his arm(s) to control pointing, as well as having a relatively long distance from the sensing device. To solve those usability problems, we propose a pointing device that can narrow the distance within which users can comfortably move their wrists between the keyboard and the pointing device itself. For this system, we compared between various depth sensors' performance and eventually chose the Intel Realsense sensor for our system's use. Additionally, we performed a comparative study involving our proposed system's performance and that of other conventional input devices. Although the total time to complete experimental tasks using our system was longer than by using other conventional input devices, our proposed system has the fastest time when switching between pointing and typing (i.e. moving hand from mouse to keyboard).

Many various robot hands have been developed so far. For grasping various objects by the robot hand stably, the grasping form based on the objects up is necessary. This paper proposes the estimation method of grasping form based on the objects with vision and tactile information for the developed robot hand. Here, the robot hand system learns the grasping form based on the objects through pick-up motion, then estimates the grasping form by using the learned estimation system.

Expectation for human symbiotic robots to support daily activities in aging societies is occurred. However, deficient sensing ability on the current robot hands makes difficult to do grasping actions and manipulation. Therefore, sufficient data and higher sensing ability are required for operating robot hands. In this research, a new robot finger mechanism is developed. This mechanism has the rotation center located at the edge of the soft skin and is able to bend finger joints by using slider, linear motor, and link mechanism. Hereby, robot finger can attach flexible skin and sensor continuously. Also, sensing ability comparison of other robot finger and the new finger with proposed mechanism is evaluated. As a result, the ability of proposed finger mechanism is confirmed.

The purpose of this study was to develop a gripper equipped with an additional robotic limb called “Third Arm” with which it is possible to perform two concurrent tasks. A gripper for “Third Arm” requires light mechanism which can grasp non-binding objects without dependence on the reaching angle to support our daily life. In this paper, we focus on adjusting shape of the object to meet required grip function and propose “Hybrid gripper” which is composed of two mechanisms. One is open-close mechanism with ball screw, the other is jamming mechanism applied with jamming transition phenomenon. We verified the reaching angle (0, 30,45, 60, 90) which the gripper can grasp objects. The angle range of hybrid gripper is wider than conventional simple gripper at each angle. The result suggests that the hybrid gripper can apply as “Third Arm” gripper.

This paper investigates tactile object recognition with relatively densely distributed force vector measurements and evaluates what kind of tactile information is beneficial for object recognition. The uSkin tactile sensors are embedded in an Allegro Hand, and provide 240 triaxial force vector measurements in total in all fingers. Active object exploration is used to gather training and testing data. A simple feedforward, a recurrent, and a convolutional neural network are used for learning. Evaluations with different number of employed measurements, static vs. time series data and force vector vs. only normal force vector measurements show that the high-dimensional information provided by the sensors is indeed beneficial. An object recognition rate of up to 95% for 20 objects was achieved.

A method of contact detection to a rigid layer in a fingertip to improve grasping performance with three-layer structure finger composed of a rubber membrane, a fluid layer, and the rigid layer is proposed. The three-layer structured finger can delicately grasp fragile objects with the rubber membrane and the fluid layer (soft grasping) and provide large resistible force with the rigid layer (tough grasping). In order to perform tough grasp with the three-layer structured finger, contact detection to rigid layer is required. If contact with the rigid body layer can be detected, it is possible to choose soft grasp or tough grasp. Moreover, by displacing the finger after an object touches the rigid layer, the amount of deformation of the object can be estimated from it of the finger.

In this study, we experimentally investigated the effect of change in contact area by deformation of soft material fingertip on the resistible force. In our previous study, harder fingertips in which small bending deformation occurs produce larger resistible forces. When a tangential force was applied, large bending deformation occurred in the soft fingertip. It is assumed that the change in the contact area by bending deformation affected the resistible force.

This paper presents a new type of leg-wheeled robot designed to reduce the degree of freedom (DOF). The main element of the design to reduce the DOF is to use a passive wheel and a 6-joints closing link mechanism. By using these elements, it is possible to construct a leg-wheeled robot with 11 DOF despite having 6 legs. In the previous studies, we verified the wheel movement. However, problems such as intense increase and decrease of movement speed occurred. Hence, we develop a prototype with only a wheel moving mechanism, devise an motion plan such as using a gliding locomotion, and consideration of the constant velocity movement.

This paper shows control algorithm for miniature motorcycle robot. Though the robot has a inertial rotor that is able to maintain roll posture, a inertial rotor is redundant degree of freedom, which deteriorates energy efficiency. Therefore, this paper proposes a method to keep balance by only steering while driving.

Our study is focusing on unmanned bike system control. In our previous studies, it is confirmed that attitude control is possible when stop and driving straight by using Control Moment Gyro (CMG). However, turning and slalom have not been confirmed. When these behaviors are performed by the bike, it is necessary to the load shift or generate the bank angle. In this study, it is considered that load shift system of unmanned bike can be performed by CMG. The load shift experiments were carried out by simply changing the control formula of CMG. As a result, it was confirmed that load shift can be performed by CMG. It means that the CMG is useful as a method of attitude control.

This paper proposes a front-wheel-driven bicycle which can maintain its balance in a stationary state. The proposed bicycle has one degree of freedom for driving the steering, and has one degree of freedom for driving the front wheel. To maintain its balance, when running, the steering angle is controlled, and when staying, the torque of the front wheel is controlled. The paper shows control laws for these motions, and experimental result shows the proposed bicycle can maintain its balance in the case of running and staying.

A 1/3 scaled four-wheeled and four-legged personal robot was designed and manufactured. Authors designed the weight reduced and higher rigid mechanism using the resin type 3D printer. Arduino microcomputer system was adopted as a control circuit. The program for the basic motions was developed and carried out.

Active textile based on three strands with thin McKibben muscles is soft, lightweight, and compliant. New knitting method of the active textile enhance a performance of the active textile owing to not using the woof and loose interaction between thin McKibben muscles. Bundling or knitting thin McKibben muscles causes physical interaction between muscle fibers with their radical expansion, resulting in the increase of the muscle contraction ratio. In this paper, a structure, a method of fabrication, and the characteristics of the active textile is reported.

The Ionic polymer metal composite (IPMC) actuator is a representative polymer actuator. Useful properties include high responsiveness, the capacity to be driven by low voltage, longevity and the ability to operate in water. However, almost all of the commercially available IPMC actuators are flat films. Because of this, most of the conventional biomimetic robots which use IPMC consist of rectangular IPMC actuators. In this paper, we manufactured a butterfly shaped IPMC actuators by a fabrication method for designable 3D shape IPMC actuators that we have proposed. The butterfly robot had several wings with different resonant frequencies and each wing was able to flap separately by sinusoidal wave voltage which included the same resonant frequency of the wing. Also the wings could flap simultaneously by combinational sinusoidal wave voltage.

We are developing 6-DOF machining robot applying new parallel manipulator system with our folding 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. Our previous model, type H manipulator is originally designed for snow-plow task and not designed for linier cylinder substitute, so the mass balance around the telescopic direction of the manipulator is shifted to the side-mounted driving motor. This mechanical unbalance also causes conflicts between each manipulator. This paper describes the development of the re-designed new telescopic manipulator type I, for the substitute of linier cylinders.

Pneumatic rubber Artificial muscle (PAM) has high back drivability and high output. However, PAM's Repetitive durability is low and not less well understood. Observing broken PAM which was made of carbon and latex rubber, there were breakage of fiber in it. Thus, we focused Aramid fiber that has high maximum shearing stress and abrasion resistance. Then, we compared load characteristics and repetitive durability between PAMs using carbon fiber and aramid fiber. Experimental result showed that load characteristic of PAM using aramid fiber is equal to PAM using carbon fiber. Furthermore, repetitive durability of PAM using Aramid fiber is 6times better than one using carbon fiber.

On the other hand, we tried to develop reparable PAM using thermoplastic elastomer (TPE). And load characteristic experiment and repetitive durability experiment were conducted. As a result, load characteristic of PAM using TPE was equal to previous PAMs. For future works, we are planning to simulate PAM from the viewpoint of material mechanics. Meanwhile, we are planning to improve repetitive durability of PAM using TPE.

Inflatable robotic arms with light weight and softness have been proposed to expand the application fields of robots. In this paper, we develop a force sensor made of a plastic sheet air-bag. Due to the structure and material, this force sensor has lightweight and flexible features. Therefore, this sensor is suitable for inflatable robotic arms. A force sensing model is proposed in this paper to estimate external force from internal pressure of the air-bag. The model is derived by measuring the contact area between the air-bag and the contact plane. Furthermore, we experimentally confirm the usefulness of the proposed sensing model.

There is a growing need for robot hands that can grip and manipulate various shape objects. In this paper, we propose a robot hand system which has a flexible actuator group and robust visual feedback control. The flexible actuator is made by a 3D printer capable of using different elastic materials and is pneumatically driven. A new visual feedback control method is effectively utilized by estimating the linear relation between the resultant force/torque and the actuator input. By using this proposed method, the manipulation of the object in a plane is realized. We experimentally confirm that the position deviation is about 0.30 mm and the orientation deviation is about 0.03 rad.

In this paper, we proposed the Universal Vacuum Gripper with active deformation (in short UVG). UVG is a vacuum adhesion cup filled with powder in elastomar membrane which adhere to uneven surface. However, the uneveness like deep step is hard to adhere by the UVG. In order to improve the adhesion performance of the step, we focused on the active deformation mechanism based on the negative pressure.In order to clarify the deformation condition, two type of active deformable adhesion gripper were developed; adhesion cup and UVG type. As the first step, we examined the adhesion cup type for clarifying the capability of the deformation based on the negative pressure. As a result, the thickness of the membrane is the key aspect for the deformation. Thus, the experiments were conducted against UVG with different thickness. We also evaluated the effect with or without powder inside the membrane.

We proposed the novel Jamming Gripper, named Contraction Universal Gripper (hereinafter referred as CUG) based on the porous granular material. The basic idea of CUG is stands on Universal jamming Gripper (UG) [1] [2], which is a robot hand that grips objects of various shapes. The grasping principles of UG are jamming transition by filling particles in an elastomer film and evacuating the inside of the film. However, conventional UG is difficult to grip a small-diameter object. Moreover UG required the strong pressing force for its deformation to the shape of the object, therefore it may leads the damage to the fragile objects.

In this research, we focused on the porous granular material. CUG based on the porous material can grasp the smaller object than UG, because it has the voids inside the material, the powder itself can be shrunk. In addition, the pressing force for the deformation may be reduced because the powder itself is softer than the hard powder. In order to clarify the characteristic of CUG, we conducted a grasping experiment, and compared its performance with the conventional UG. In addition, the relationship between pressing force and gripping force is evaluated.

This paper proposes a concept of the task-oriented software synergy (TOSS) as a novel control method for anthropomorphic hands and a control system of multi-fingered hand that realizes TOSS. TOSS is able to perform various tasks with fewer inputs than conventional synergy by switching synergy for each task. By comparing TOSS with the conventional synergy, we show that TOSS gained higher contribution rate than the conventional synergy for writing task and smaller error of joint angle due to the effect of dimensional reduction. Moreover, we verified through physics simulation that the success rate of a task using TOSS with a few principal components is higher than conventional synergy based method. These results show that TOSS can properly perform various tasks with fewer inputs.

In the robotic grasping, occlusion occurs around target object by surrounding objects when using vision sensor. In this case, it may cause the grasping failure due to the collision by the surrounding obstacles. Therefore, In the robotic grasping, In order to solve this problem, we propose a method by installing proximity sensors on fingertip and nail besides on finger belly. These sensors can detect the position and distance information to the surrounding objects for avoiding collision. The proposed method has been implemented to a robot hand system and successful avoiding motion has shown the effectiveness of the method.

Japanese joinery is one of the key technologies to build Japanese wooden architecture. The joinery is used to connect wooden beams without using nails and glue. We use this Japanese joinery combined with 3D printer technology to develop robot fingers. This method simplifies assembly process, and decreases the volume of the robotic fingers.

This paper describes grip control which use image recognition by a robot hand applied with iris mechanism. We developed a robot hand applied an iris mechanism for expand gripped targets. This robot hand has six blades, and can be opened and closed with a single actuator. The blades can be gripped all around a target. We also developed a robot arm for this robot hand as an end effector. The robot arm is a SCARA type, and moves automatically the hand position by image recognition.

When children undergo medical examinations and surgical operations at hospitals, they sometimes need to be held down to avoid risk of injury caused by their unwanted motions. Fixing the body to the bed places stress on the children may seem inappropriate to their guardians. To solve this problem, this study developed a hoodie that restrains the body by variable stiffness elements using particle jamming only when needed. Fixing devices that use particle jamming currently in the marketplace require changing the shape by hand to fit them around the body. The developed hoodie changes shape to fit the body and restrains body motions only by evacuating the inside air.

The universal gripper has attracted attention due to its simple structure and advanced grasping ability of irregularly shaped objects. In this research, we propose a novel design for a granular-jamming-based gripper, which uses a transparent filling and semi-transparent membrane to perform optical sensing to detect both the deformation and the object that is being grasped. By adjusting the refractive index of an oil mixture to the refractive index of the granular bodies, we achieve a fully transparent filling that allows the use of a camera inside the universal gripper. We present the materials and development of our prototype and confirm experimentally that it can successfully acquire tactile information.

Contact type end effector causes several problems to the object, such as the physical damage, static electricity, metal contamination, and so on. These problems adversely affect semiconductor wafers and glass substrates. Therefore, non-contact type adsorption are focused. The application of the non-contact type is limited to a flat workpiece, and it's difficult to grasp a workpiece having an uneven shape. In this research, we propose the novel non-contact type vacuum pad, named Universal Vacuum Gripper (in short, UVG) based on Universal Jamming Gripper [1] (hereinafter referred as UG). UG is the deformable robotic gripper based on Jamming transition of granular material. UVG realize the non-contact gripping to the workpiece with uneven surface, because the lip part of the gripper consists of UG, which freely deforms with respect to the workpiece.

Soft robotic fingers have attracted a lot of attention in recent years because their safe interaction with human. However, soft fingers cannot resist large external force due to the softness of the material and complaint of the structure, which can limit their applicability. This paper presents the development of passive elements with variable stiffiness for soft robotic finger based on multi-material 3D printing and layer jamming. Both the soft-bodied robotic finger and hard stiffiness variation unit are directly printed out by a multi-material 3D printer, Object350Connex, without additional casting process. Layer jamming is used in this paper to stiffien the proposed finger. Compared to the initial state, the finger’s stiffiness is increased by up to 3.27 times when the inflation pressure is 30 kPa and the vacuum pressure is -20 kPa. The proposed stiffiness variation method has prove its effiectiveness in stiffien soft finger though a small, flexible and light passive element.

There are various foods such as fried chicken and hamburger in a box lunch. Foods have large variations in shapes and contain a lot of flexible objects. Therefore, it is difficult to grasp with a robot hand. In this research, we made a robot hand which grips objects based on binding. This hand grasps by surrounding the object with elastic yarn. We propose a mechanism in which the opening and closing of the hand and the winding of the binding yarn are linked. We also examine the relationship between the tension of the binding yarn and the force applied to the object.

To attach and detach permanent magnets with an operation force smaller than their attractive force, Internally-Balanced Magnetic Unit (IB Magnet) has been developed. The unit utilizes a nonlinear spring with an inverse characteristic of magnetic attraction to produce a balancing force for canceling the internal force applied on the magnet. This paper extends the concept of shifting the equilibrium point of a system with a small operation force to linear systems such as conventional springs. Aligning a linear system and its inverse characteristic spring in series enables a mechanism to convert displacement into force generated by a spring with theoretically zero operation force.

To verify the proposed principle, the authors realized a prototype model of inverse characteristic linear spring with an uncircular pulley. Experiments showed that the generating force of a linear spring can be controlled by a small and steady operation force.

Human hands have multi-joint multi-degree of freedom structure which makes it possible to achieve dexterous handing. On the other hand, robot hands are less functional and less versatile than human hands, and then can not fil in for human handling. The purpose of this paper is to analyze the movements of hands in various tasks performed by human and to clarify the minimum number of variables required for reproducing the movements of the human hands. We use principal component analysis and a simple feedforward neural network with three layers for the analysis. The results finally show the possibility of opening the way to a low-cost robot hand which is simple in structure although it is sufficient to achieve dexterous handling like human hands.

Small humanoid robots are expected to advance into society in the field of nursing care and welfare. Among them, the home robots need to move freely even in space with obstacles. Based on these facts, we have the goal to let climb between walls on a small humanoid robot. By the previous work, we made a small humanoid robot and the robot succeeded in climbing between walls. I improve the stability of this developed action using theory of grip. Using this theory, I calculate the required joint torques and the optimal position to support the robot. Then I compare the results obtained with the current results. In addition, I verify the maximum raising distance per cycle.

This paper presents a novel wheel that features dynamic braking of the rollers of an omni-wheel. This results in interesting motions that would not be attained by ordinary wheels or the typical omni-wheel especially for dual differential drive robots. The design and prototype of the wheel is presented together with basic modeling.

When a rescue robot is active at a disaster site, actions that cross the valley may be necessary. I thought about developing a robot that uses bridges to safely cross the valley. A bridge mobile robot uses a crawler to move in a flat place and a chain bridge to cross a valley. The robot can retrieve the bridge after it finishes crossing and uses a bridge again. This research makes it possible to expand the range of activities of robots at disaster sites beyond now. I made a model of a holder to store the chain and made a proposal for a motion to bridge the bridge.