Operation of remotely operated underwater robot (ROUV) is known to be difficult due to limited sensor information from the robot. We proposed to apply ta haptic shared control (HSC) approach, in which a human operator and an automation interact congruently in a perception-action cycle to perform a dynamic task by force inputs of both agents to a single operation input terminal., to the ROUV to increase operability of the robot.. We have prototyped an underwater robot with the HSC to investigate the effectiveness of our proposed method. Subjects were instructed to operate the underwater robot to the goal along a desired path line. An experimental factor was control method, which includes 1)HSC, 2)Fully manual, and 3)Fully automated. The subjects could partially see the desired path line and a goal visual target measured by a camera attached to the robot to operate the robot. In the HSC condition, the subjects could feel the force exerted on the joystick by the automatic control, which guided the robot to the desired path. Experimental results showed that HSC achieved lower mental workload than manual control. In addition, the HSC achieved high task performance by means of shorter task completion time than that in the manual control, while there were no significant difference in the tracking error between shared and manual control.

We have proposed a method of a Resolved Acceleration Control (RAC) for Underwater Vehicle-Manipulator Systems (UVMS) which is a position control method and a position-based impedance control method with the RAC for UVMS. In this paper, we show the usefulness of the proposed control methods by an experiment using a 3-link dual arm underwater robot that grip a fixed underwater object.

The needs of marine observation systems have been increased to make fishing industries more efficiently. However, previous marine observation systems were expensive and difficult to handle them. Therefore, the module type marine observation system has been developed. Each module has different function, such as camera and CTD sensor module. Combining different modules, the module type marine observation system will adopt to various environments. Since the module type marine observation system can maintain separately, the cost of maintenance can be suppressed. In this paper, the software of module type marine observation systems has been also considered. Since, the Node-RED program tool handles the IO as a module, it makes possible to modularize the program.

In this study, a small glider type underwater robot for environmental surveying is being developed. It is expected that the introduction of the glider system will extend the survey range of the robot by several folds. In order to evaluate the optimal hydrofoil shape, energy efficiency and motion of the robot, wing simulation is necessary. This paper describes the steady and non-steady state wing characteristics and simulation using a flat plate wing. The authors measured the underwater characteristics of a flat plate wing and focused on the acceleration terms especially the added mass effect due to its strong influence on acceleration. Added mass was measured and modeled, and introduced into the simulation. As a result of the introduction of the added mass into the simulation, the stability of the model was improved.

This study aims to develop a small glider type UROV for sampling bottom sediment of lakes. A glider system is expected to extend the survey range of the robot. To achieve efficient repeated diving and surfacing, a lightweight and small buoyancy controller is necessary. In this paper, a buoyancy controller that uses a carbon dioxide cylinder is developed. Basic discussion regarding design and verification experiments are presented.

This manuscript describes a tensegrity-type underwater robot with a transforming mechanism. The robotic body is composed of tensegrity structure with plastic pipes (struts) and a rubber band (deformable strings) so that the robotic body has deformability. A winding unit attaching to one tip of the tensegrity structure can transform the body of the robot. The winding unit includes a planetary gear. Winding the gut fixed on the rotary shaft of the water-proofing motor transforms the shape of the deformable tensegrity structure. In this manuscript, to evaluate the performance of the transforming mechanism, we investigate the fluid characteristics of the robot against constant water flow. Based on results, we find that transforming the robot with the winding unit alters its translational performances because the projecting area decreased for transforming the robotic body.

We are developing a sensing device that can keep a fixed position and move to arbitrary position on water for long term environmental measurement. In this paper, we propose a localization method of the sensing device and show experimental results of its movement. The proposed localization method is aiming to move to arbitrary directions and keep its positions automatically on water. From the several experiments, we confirmed the performance of the device.

Miniature aquatic robots are required for survey in a narrow space. In this study, we have developed a robot with multiple elastic plates which are selectively resonated by a single vibration motor. In this paper, a bird’s eye view camera is used to detect the position of the robot with a marker and the motion control is applied to let the robot reach the designated position.

We developed AUV to realize phototaxis movement of single-celled organism, Chlamydomonas. Chlamydomonas moves spirally and reach to a light source by using only one optical sensor. Our developed AUV has only one actuator and one optical sensor based on phototaxis mechanism. The AUV is composed of a sealed conatiner, a screw, and an exterior. The sealed conatiner contains battery and motor. A shape of the screw is asymmetry to realize spiral movement. The exterior makes water flow. An experiment shows that our AUV has an ability to float with spiral movement.]

The Manta Law, one of the surveys of coral distribution areas, conducts surveys visually while people are being towed by ships, which has a big burden on the human body and is not suitable for long-term investigation.

This laboratory has been developing a towing type underwater robot for the purpose of reducing the burden on the human body. A towing-type underwater robot under development is equipped with a camera and attempts to acquire images.

In this paper, we perform altitude measurements from the ocean floor using ultrasonic wave to maintaining the altitude of the towing type underwater robot for image acquisition.

Water supply system is essential social infrastructure in these days. Slow sand filtration method which is a type of water purification method has the advantages of less environmental stress and lower cost for long-term use. The method has two disadvantages. One is needs of constant cleaning of a bottom of a pond, which conducts every 2 or 3 months because of its clogging. The other is difficulty of cleaning at coldest period because a pond is covered with ice. We have been developing an underwater cleaning robot for slow sand filtration pond to eliminate the disadvantages. In this paper, methods of stirring deposit and moving at severe cold period are reported.

We focus on high mobility underwater locomotion peculiar to fish, and study on a fish-type robot with snap-through buckling. In this paper, we aim to increase the swimming velocity of the robot by improving the frequency at which snap-through buckling occurs. We propose a new driving system of snap-through buckling with slider-crank mechanism. We developed a robot with the proposed mechanism, and measured the frequency at which jumping buckling occurs and the swimming velocity. From the results, we achieved increasing the frequency and the swimming velocity of the robot. Moreover, we fined that the relationship between the frequency and the swimming velocity has a single peak.

This paper discusses the development of exploration robot inside containment vessel aiming at the decommissioning of Fukushima Daiichi Nuclear Power Plant. Focusing on ultrasonic measurement, the developed robot supports transportation and measurement of the measuring instrument. In order to achieve this, the robot has the landing function for stationary and the underwater moving function. In order to integrate these two functions, we designed and prototyped, examined optimization, and verified its effectiveness by control experiment.

This paper demonstrates a whole-body compliant motion by using resolved acceleration control, on the hydrostatically driven humanoid Hydra. Although Hydra realized a compliant and robust motion control in a previous study, it was limited on the ideal planar surface. In this paper, a robust standing balance on an inclining seesaw is achieved, utilizing a compliant motion realized by resolved acceleration control and a state estimation by IMU sensor. Moreover, an experimental comparison with joint position control and resolved viscoelasiticity control is shown, about the biped walking performance.

Shape Memory Alloy (SMA) is an excellent actuator for small robots due to its miniaturization adaptability and large deformation. By applying SMA to joints of small walking robots, mobility of the robots can be expanded. The response characteristics of SMA depend on its phase change by heating / cooling. In the SMA heating, its response time becomes faster by supplying higher heating power. In contrast when the SMA cooling, a forced cooling mechanism is required for becoming the response time faster.In this study, we propose "the movable heat sink cooling mechanism" as a compact mechanism which can achieve the arbitrary control of the cooling timing. The proposed mechanism was applied to a walking robot whose joints were driven by SMA, and it was demonstrated that its response characteristics were improved.

This research report describes a model of cheetah ’s flexible spine and hind leg by utilizing a link mechanism. The mechanism consists of a number of rigid links, which is driven by two pneumatic actuators. One actuator drives the link mechanism is for reproducing the curvature of the cheetah ’s spine during running, and the other retracts the hind leg to generate clearance. We have developed the smallest set of the spine and a hind leg so that we could demonstrate simple running behavior.

In this paper, a new walk control for a hexapod robot to walk on irregular terrain with an operator’s foot-placement navigation is proposed. The control method is based on Follow-the-Contact-Point (FCP) gait control which has a principle that each leg just follows the contact point of its foreleg. Consequently, planning the contact points of all legs is summarized to one of the front legs. In order to realize adaptive gait for a hexapod robot, three control architectures that consist of walking, planning, and local adaptation are proposed. First, a real-time posture control and new constraints in transition from a stance phase to a swing phase are added to keep static stability when a leg leaves the ground. Second, a planning method based on the advantage of the FCP gait control is added. Third is an adaptive control to adjust the time elapsed in each control mode, by which specifications of deadlock-free and static stability are satisfied. The proposed control architecture is installed to a small hexapod robot, and whose validity is verified through experiments in which the robot walks on uneven terrain.

When a quadruped robot walks, there are two kinds of walking methods which are dynamic walking and static walking. Either of the walking modes cannot keep their trunk constant-posture/velocity at high-speed without a special mechanism. This is caused by the large inertial force generated from legs of the robot moving at high-speed. We propose a method to cancel inertial translational force by attaching a counterweight to the upper side of each leg and set gaits and leg trajectory appropriately so that the trunk keeps constant-posture/velocity in high-speed walking. Furthermore, as a result of kinetic analysis and actual machine experiments of the counterweight quadruped robot, we showed that it is possible to suppress the influence of the inertial translational force.

We have confirmed that passive elastic trunk joints affect the energy efficiency of quadruped robots. The main causes of energy loss are negative work and contact between the legs and ground. We clarified relationship between legs and trunk joints for decreasing energy loss. Elastic energy of the trunk joints increases at the timing when kinetic energy in robots decreases. On the other hand, elastic energy of the trunk joints is released at the timing when kinetic energy in robots increases. Energy loss is smaller by this mechanism. In this paper, we develop a quadruped walking robot having passive elastic joints in the trunk. Also, by this walking robot, we verify that the proposed relationship between leg and trunk joints is reliable.

Research on a walking robot focusing on the movement of waist on the left and right of the lizard. In the case of walking using only the lumbar region, slippage occurs in the foot. In order to clarify the effect of foot slippage on walking, the relation between static friction coefficient and moving distance was investigated.

In order to investigate the influence of foot translational motion on walking, we studied a waist driven walking robot with two Yaw axes. Numerical calculation was carried out for the angle of the Yaw axis which can make the distance between supporting legs constant. For the initial angles of 30 ° and 20 °, the walking cycle was changed and the distance of one step was measured.

Recently conflicts between wildlife and human becomes worldwide problem, but conservation of wild animals is also a current hot topic. We are developing a unique multilegged robot“WhiteGoat” series for animal menacing to reduce wildlife damage. For the goat series, we employed simple mechanical leg driving system - inverted Chebyshev link with 3D trot vectoring mechanism, so the durability is relative higher than conventional legged robot. We have already reported the desk top sized prototype, but the size should be not enough for menacing bears etc. In this report, we will describe the next generation of the robot series, goat-size prototype “GuardianGoat”: the expanded driving mechanism and telescopic manipulator as non-lethal threating equipment.

In this research, we developed AeroCat, the quadruped robot with pneumatic cylinders and electric motors combinedly. In order to save lives in irregular places such as disaster sites, AeroCat explores by jumping and landing ability. Conventional rescue robots has a linear actuator inside a leg links, such as a pneumatic cylinder. Generally, pneumatic cylinders are difficult to position control, thus it is difficult to perform walking motion using only pneumatic cylinders. Besides, the size of the built-in actuator was limited, that lead the limitation of the jumping and landing ability. In this paper, we propose the hybrid mechanism with pneumatic cylinder and electric motor based on knee follower mechanism. The knee follower mechanism is transfer the movement of the large cylinder outside the leg links to the knee portion to achieve the high jumping and landing. In order to realize walking motion, an electric motor is installed in the hip joint to control the angle of the leg. We evaluate the moving ability of the quadruped walking robot with the proposed leg.

This paper proposes a walking control design for the utilization of passive elastic trunk and shows the result. It is inferred that legged animals decrease negative work during walk by utilizing elastic elements: muscle and tendon. There are researches on legged robots by utilizing elastic elements as the elastic utilization of legged animals . However, there are no researches on the walking control design of the legged robot for elastic utilization. Therefore, we aim to achieve to decrease energy loss of the walk of a quadruped robot with a passive trunk.

Researches on biomimetic robots that realize movement with fewer actuators have been conducted. In this research, we change the linkage mechanism of the lizard-type robot of the previous research, and investigate the motion in a plane surface driven by single actuator. In this paper, we explain the mechanism of walking robot and walking algorithm, and perform modeling and simulations based on kinematics.

Humanoid robots have potensials to walk on uneven ground unlike wheeled robots. However, it is difficult to reach a given destination without falling down based on only visual sensors. To reach the destination safely, we propose the autonomous safe locomotion system applying vision and sole reaction force to the footstep planning. Moreover, the robot reuses the results of the motion by saving them in the database with the foothold label given by the visual classifier. Experimental results using a real bipedal robot CHIDORI with the propsed system are shown at last.

Recently, powered exoskeletons are researched and developed as a device that assist with human movement, and solution for labor shortage, in the field of agriculture, nursing care and medical. We study and develop the powered exoskeleton that assist decontamination workers at nuclear power plants in a disaster. The latter propose feedback control by force sensors and feedforward control of joint angle trajectory based on LPC (linear predictive coding) for powered exoskeleton’s posture angles. Each foot sole is equipped with a force sensor. It is easy to measure the force acting between the powered exoskeleton and the wearer when moving the legs because the exoskeleton does not band the wear’s legs with its legs. LPC based feedforward controller enables the powered exoskeleton to predict the future joint angles with small amount of data and assist the wearer’s walk more smoothly. We conducted experiments and the results show the validity of the proposed method.

A powered exoskeleton has attracted attention as a solution to a shortage of labor in fields of nursing care and agriculture. Electromyographs and foot switches are often used to control the powered exoskeleton. We have been developing a powered exoskeleton to support workers at a nuclear power plant in case of a hazard. In this case, electromyographs are not suitable because it is difficult to measure electromyographs because of sweating caused by high temperature and humidity. The foot switches are also not suitable because they require two or more steps of motion for the walk recognition, so it is difficult to promptly assist. We have proposed a power assist control method based on motion prediction using 9-axis motion sensors solving these problems. The conventional method needs five motion sensors attached to the torso and the lower limb for movement measurement. The control method based on some machine learning techniques requires a lot of teaching data and it is difficult to learn in real time. This report proposes to use one motion sensor on the wearer’s torso for motion recognition. Experimental results show the validity of the proposed method.

In this study, in order to control the position of a human arm to the target trajectories or surfaces in cooperation with a human motion, we developed a 5-DOF skill-assist arm with passive joints. We consider that the skill-assist arm can be applied to a forearm rehabilitation and so on. The skill-assist arm consists of a 3-DOF serial arm that can control a position and a 2-DOF rotational mechanism installed at a tip of the serial arm. In this paper, we proposed control methods of aiding a position of the human arm to target virtual planes and free curved surfaces in space using the skill-assist arm. In order to aiding the position of the human arm in cooperation with the human motion, the skill-assist arm switches passive or active states of the actuators of the 3-DOF serial arm according to a direction of the target surface and an attitude of the skill-assist arm. After proposing the control methods, we conducted experiments using a prototype of the skill-assist arm and we confirmed an availability of the control methods from the results of its experiments.

The purpose of this study was to develop an additional and remotely controllable robotic limb called “Detachable Body” which enable perform two concurrent tasks. In this paper, we studied the video presentation method to enable perform two concurrent tasks in a remote work area using “Detachable Body”. We made it possible by overlapping two images. Therefore, we comparison of preferable transmittance when images were layered. The transmittance was verified in three ways from the viewpoint of attention distribution for work. As a result, it was suggested that the transmission ratio suitable for distribution / movement of attention is 3: 7 or less in one use case.

We aim to build a Brain Machine Interface using motion and image recall for manipulation of auxiliary robot arms. The development of auxiliary robot arms has made good of the labor force drop due to the lack of labor.Many current auxiliary robot arms are designed to follow hand and foot motion of the operator, so that the operation of the robot arm is restricted under the body motion. By using BMI, it is possible to operate an auxiliary robot arm with high degree of freedom. therfore, we utilize EEG when the user recolle of grasping motion of hand and robot movement image. Selection of image task and describe the construction of a system to process brain waves in real time and runs robot arms.

Cart control requires propulsion and steering simultaneously. It will impose the user a hard workload in the lots of human environment. To overcome this problem, we design a safe auto-steering mobile system. The steering is controlled by an attractor controller that entrains the cart to the reference trajectory based on vector field. The trajectory and vector field are designed under non-holonomic constraints, and the controller is designed by functional approximation with an averaging method of the vector field. A cart system is designed and prototyped. The experimental results show that the propulsion force obtained from the user confirms the safety against collision with pedestrian.

This research is to study the effectiveness of force communication between the driver and the proposed vehicle, while driving. In the proposed system, not only the driver indicates his/her intention to the vehicle by rotating the steering wheel, but the car indicates its intention to the driver by force feedback via the steering wheel. The car makes sequence models from the past driving sequences under each environment. If the vehicle judges that the driver will trace one of the sequence models, it lets the driver to trace the sequence by the force feedback via the steering wheel. We simulate this experiment with a computer. We have studied it using an operating sequence model. But the vehicle trajectories often have complete difference, even if the operating sequences have a little differenece. In this paper, we make sequence models from vehicle trajectories based on environment coordinate.

We propose a new steering method in a case of steer-by-wire vehicle. In general, A steer-by-wire system is designed so that the rudder angle of a vehicle becomes the angle according to the steering wheel angle as soon as possible. But we research human tolerance against the difference between the radar angle and the angle with respect to the steering wheel. And we propose the method to reduce tire wear and to reduce energy consumption within range of the human tolerance. In this report, we research the human tolerance corresponding to movement of the vehicle on the actual vehicle(costmized coms), and we propose the steering method with considering the torealance.

In this study, the maneuverability of a control method for a power-assisted cart is evaluated using the operator’s moving velocity. In the past study, we proposed the control method using control point that is the position to keep the velocity on the cart coordination. The operator’s relative velocity is calculated by the data of LRF on the cart, and the walking speed of the operator is estimated. From the experimental results, we confirmed that the operator walks constant speed during the operation when control point was set nearby the operator’s standing position.

For human work, there is a pair of the force and velocity that maximizes human power. From this viewpoint, power assistive system has to be designed to assist both the force and velocity. So far, several systems dealing with both of those have been proposed, the assist force was determined heuristically. In this paper, we propose a force and velocity assistive system based on characteristic of human muscle. A pair of the force and velocity, which maximizes human power, is calculated from the relationship between the human force and velocity. We design an assistive mechanism that maximizes human power when he rotates a handle. In the proposed system, the force and velocity are adjusted independently by two motors connecting via a differential gear. We evaluate the power assist performance through the experiments.

The pico-EV・Eco Challenge is a competition in which the vehicles which carried a person run during 20 minutes using a low powered battery (7.2 V, approximately 1000mAh). This competition has been held after 2012 every year. In this paper, these competitions held until now will be outlined. And, about the pico-EV Eco Challenge 2019, after describing the content of this competition, the motor, the control techniques, the reduction ratio, the body specifications, the results of the competition, etc about the vehicles will be shown.

There is a concern about the food crisis due to worldwide population increase and the future rise in food prices. It is therefore preferable to realize local food production for local consumption. We are developing a small indoor hydroponic system for household use, which uses sustainable energies, such as solar, wind, and hydrogen power. We have conducted experiments using a small hydroponic system to determine the environmental situation of the lettuce cultivation process. This paper will explain the developed low cost environmental sensing system and the results of the vegetable growth process with LED lighting control using small indoor hydroponic system.

In recent years, not only various driving assist systems for improving safety for humans are introduced but driving assist systems to improve ease of driving and comfort are also progressing. Since these systems assist all kind of driving behavior, it is necessary not only to design a single vehicle system but human and automobile systems are also needed. This research aims to develop an assist system that helps beginners and elderly to adapt in a short time. By using neural network to imitate the characteristics of the human behavior, this research optimize the dynamics of the steering target tracking error.

Push carts have been widely used in various areas such as factories, markets and buildings around the world. However, we are frequently faced with the difficult situation not to move the push cart because of the heavy load mounted on its cart. For this reason, many assisting devices to physically aid the cart operator have been developed, and some of them have also been commercialized. On the other hand, these former assisting devices have troublesomeness of manual switching of with/without assisting or direct furnishing and unfurnishing, and it is additionally difficult for most former devices to assist the turning motion of the push cart. Here we propose a new assisting system for turning a push cart without manual troublesomeness.

A pressure distribution sheet sensor (LL sensor) by electro-magnetic coupling and 8 pieces of resistive pressure sensor (R sensor) are used to convenient estimate human movement. Center of pressure (COP) is calculated by the pressure data of 37 x 45 sensor matrix. A subject stand on the LL sensor with putting on slippers those are with Linear regression is adopted in this study with scikit-learn software package. The predicted values are compared with two calculations, 8-point data of LL sensors and of R-sensor. While total prediction error is almost 4%, that under the static posture is almost 1%. This seems to be ascribed to the sensor characteristics (linearity and time responsibility) and its detection area difference.

The analysis and modeling of driving behaviors are important to design a comfortable functions for the driving systems. The estimation method of the steering model using the eye tracking information has been proposed using the heuristic search algorithm. In this paper, we proposed hybrid steering model. The system is switching from normal driving into inattentive driving according to the field of vision. Using the gazing point measured by the eye tracking device, the non-linear driver’s steering model that minimizes the error between the vehicle and its model is estimated by Particle Swam Optimization (PSO). However, in case of an inattentive driving, it is necessary to estimate the model robustly in realistic conditions.

Japan is facing the problem of super aging society and the necessity of assistive system for the elderly has increased. In order to decrease the human resource and cost, we need to develop the technology to predict and support human motion. Conventional motion prediction system uses force plate or motion capture but these system is limited by the environment so we focus on the Electromyography(EMG) which is available as portable sensors. Our motion is generated by the force and the force is generated by our muscle. Therefore, we consider we may predict our motion from muscle activity, which is the origin of our motion.

In this research, as a first step to realize the better motion prediction system, we try to predict ground reaction force(GRF) from EMG using Recurrent Neural Network.

People counting provides valuable information on evaluating an event, marketing, and security. Therefore, it has been proposed that the systems count passing people by using a top-view camera installed at a high position in general. Here, we take into account that it can be easily installed anywhere, we propose a portable system for counting passing people by using a side view camera with a fisheye lens. In order to count passing people, we used a background subtraction method to detect moving region, and calculated its vertex. Furthermore, the direction of passing people was estimated by using optical flow. In the verification experiment, it was confirmed that the developed system could correctly recognize a walking direction and count the number of people.

In this report, biometric information such as passenger's body temperature, metabolic rate, etc. due to difference in air temperature and volume are acquired for the automobile air conditioning system. Therefore, basic knowledge about the air temperature and volume which humans feel comfortable was obtained.

Patients of neural intractable diseases should be able to communicate easily. However, patients could not continue to gaze at one point, and it’s difficult for them to use the devices which cause users to calibrate. In this study, we developed a line-of-sight input system without calibration and contact. This system needs no calibration and could respond to the users’ posture misregistrations, owing to installing the two infrared LEDs in the both ends of the monitor and the two USB cameras in the lower of the monitor. The line-of-sight direction could be estimated by the relative position between pupils and corneal surface images detected by corneal light reflex method.

In this study, we constructed a system to estimate position for the tips of human`s toes in the dark environment by using single infrared camera. Here, we recognized the tips of human`s toes presence by binarization procedure and distinguished between a target toe and background using a difference method. In addition, we applied this system to the movable robot that can obstruct the wandering elderly person. The robot firstly found the legs of a wonderer and searched the tips of the toes. After that, it approached a target person to maintain a constant range. As a result, we could construct a system that estimates a target toe position for a watching robot.

In this study, we constructed the recognition system for nostril position based on skin color information. Here, rough information of the original image was used, and mosaic processing was applied to the image. For an obtained image of 640 × 480 pixels, 16 × 16 pixels were translated into one block and obtained image was divided into 40 × 30 blocks, and the representative color of each block was obtained from the average color of the representative points. Blocks whose representative colors satisfy the condition of the skin color were extracted and smoothed further to estimate the rough face area. In the measurement, there were many dark areas around the eyes, mouth, and jaws, but the area around the nostrils had a high proportion of bright skin, and the area around the nostrils was estimated by finding the proportion of skin color in the face area. This characteristic was used to confirm the detected nostril position, and it could prevent misdetection of nostril positions.

Japan is about to enter a super aging society. The autonomous movement of the elderly is important, and the use of personal mobility vehicles such as wheelchairs is increasing. If the elderly travels through the crowd using a wheelchair like a supermarket, there is a possibility that a wheelchair user may become a psychological burden due to the relationship with surrounding people. In this study, we investigated by using visual field information entering the eyes of a wheelchair user or using skin potential reflection (SPR).

Car seats are accompanied by an adjustment mechanism that allows comfortable posture. The increase of female drivers is remarkable, the muscle strength tends to be lower than men, and it is necessary to find the lever position which is particularly easy to manipulate. In this study, we used a precision jig on the side of the seat and measured subjective and EMG integral values at multiple positions on the adjustment lever. The EMG integral value with large individual differences was analyzed by mixed effect modeling in which subjects were variable effect and lever position was fixed effect. As a comfortable position of lever to manipulate for both males and females, the existence, position, where the subjective evaluation was good and the EMG integral value was low was statistically significant.