The more joints a robot arm has, the more calculations are required. If one tries to make a humanoid robot, the amount of computation for control becomes enormous and control becomes difficult. Therefore, we propose a control theory using exponential coordinates. We believe that the use of exponential coordinates will facilitate the control of tree-type robots, which has been considered difficult with conventional control methods. To demonstrate the above theory using exponential coordinates, we will design a robot finger that can control each joint.

In recent years, master-slave system equipped with multi-fingered hands that can be operated intuitively and intricately in hazardous environments, such as disaster sites, are in demand. However, there are two main problems. First, human hand movements have a high degree of freedom, making it difficult to accurately measure shapes in real time. Second, because of the kinematic structural differences between the human hand and the robot hand, it is difficult for the slave side to properly represent the motion intended by the master side. To solve these problems, we propose a method for mapping via grasping features and compensating the mapping using grasping motion classification. Experimental results show that the proposed method enables smooth teleoperation of the robot hand, and the introduction of SVM improves the accuracy of grasping motion classification.

Covering robot fingers with thick tactile sensor skin increases the adaptability of the hand for various tasks. However, covering certain parts of the hand such as the joint with sensors is challenging. This paper presents a novel structure to cover a robot finger with thick tactile sensor skin. By encircling the joint and finger phalanges with the skin, the robot finger can have a vast sensing area on both the palmar and dorsal side of the finger, including the joints at the same time. uSkin tactile sensors, which are distributed 3-axis tactile sensors for flat and curved surfaces, are implemented for evaluation. To validate the novel structure, we confirmed that continuous tactile sensing on both sides of each joint is possible. Finally, this paper also shows that tactile sensing on the palmar and dorsal sides can be useful to improve manipulation stability.

In recent years, soft robotic hands that mimic the functions of human hands have been studied. While grasping performance can be improved by adding the characteristics of human skin, the number of parts increases when trying to reproduce multiple degrees of freedom, causing problems in assembly and maintenance. Therefore, we aimed to create a robot hand that has the advantage of reproducing the human body and is easy to assemble by integrally molding the skin and skeleton into a single part. In this study, we describe the design and fabrication methods, and show that a 15-DOF robot hand can be fabricated in less than one hour.

Point clouds describing fingertip workspace can be used to generate precision grasp poses with no constraints on the robotic hand structure. As during generation, the k-d tree for neighbour points searching between object cloud and workspace clouds is frequently used, and also because the O(logn) averaged searching time complexity is highly based on the number of points inside the cloud, how to reduce the number of points inside each workspace cloud by removing invalid points should be one way for the acceleration of grasp pose generation. In this paper, we propose a method to process workspace clouds based on the analysis of target grasp object and also the middle points of fingertips contacting. Experiments have been done in conditions of parallel gripper, 3-fingered and 5-fingered robotic hands.

In recent years, there are many researchs related robotic hands. The most important ability on robotic hand was accurate and quick repeatable movements so far. Therefore, it was enough with gripper type as a robotic hand shape in most cases. But most of human-like robotic hands are developed for specific purposes not for general purpose. In the previous study, Introduces Human-like 3-DOF Robotic Finger using mechanism named Twisted round-belt actuator which yields contraction forces produced by twisting an elastic round-belt. Accordingly, In this paper, we introduce an optimal fingertip force control method for a three ‐ fingered robot with a newly designed module(VpSM). This module can eliminate the irregular torsion phenomenon of the circular actuator (TbA). Experimental results of tracking control of step force and sinusoidal force are presented. We also demonstrated the optimal fingertip force control method for 1 Hz, 3 Hz and 5 Hz sine wave.

This study deals with a cooperative route guidance for a ground mobile robot with visually assisted by an autonomous drone. The cooperation between the ground robot and drone is generated within a symbolic motion plannar using behavior tree. The drone observing the ground envrionment generates the occupancy grid map to find the passable routes. The map infromation is shared with the ground robot which also configures the occupancy grid map based on the its own ovservation, and both of the map is integrated to find the optimal route for the ground robot. Experimental results illustrate the feasibility of cooperative navigation strategy.

Recently, cooperative transport systems using multiple robots have been attracting attention. Multiple robots can flexibly form a formation according to the task, and the robots can cooperate to efficiently transport the object. Cooperative transport with formation change consists of 5 tasks: (1)grasping the object to be transported, (2)transporting, (3)releasing, (4)changing the formation, and (5)grasping a new object to be transported. The formation is changed with the appropriate sequence and timing of these actions. In this paper, we develop a control system using state machine for cooperative transport by using a formation change of robots, a grasping/releasing for the objects and robust transportation proposed in previous studies. A sequence of operations for a cooperative transport system is managed by the state machine with 5 tasks. In the experiment, the state machine is configured using smach, and the effectiveness of the system is evaluated using an experiment of the cooperative transport with formation change.

Multi-agent Pickup and Delivery (MAPD) is a problem to compute non-colliding paths for multiple mobile agents to complete a stream of delivery tasks. MAPD models the actual delivery systems like automated warehouse, on-demand bus, and so on. Unlike existing approaches which solve the problem by a central computer or a set of autonomous agents, we solve MAPD problem based on a concept of active environment: the infrastructure assisted navigation system with spatially distributed computing units. By this concept, we aim to make a scalable system without single point of failure, while taking over the advantage of centralized systems. In this paper, we develop a distributed algorithm extending TSWAP, an online method for MAPD, and give a proof that the algorithm eventually completes all tasks. We also evaluate the proposed algorithm by a computer simulation.

The authors have been studying a welding robot system with the cooperative towing of a group of robots. This paper proposes a method for guiding a towing robot in a gentle slope environment, where the influence of the vehicle’s gravity and the constraint force of the cable fixed to the towing vehicle must be considered.

We proposed a robotic hand using elastic sheets and openable two-fingered hand. This hand can grasp the objects by opening the fingers of robotic hand and pulling the elastic sheets. Because the sheets are flexible, the hand can grasp non-uniform shaped object. The grasping force is proportional to the tension of the elastic sheets. Therefore, the weight of object is larger, the grasping force is larger. We measured the grasping force of prototyped robotic hand. The maximum force was 50 N when the diameter of cylindrical object is 70 mm and displacement of robotic hand was 50 mm. When the diameter of cylindrical object is 40 mm and displacement of robotic hand was 24 mm, the elastic sheets slipped on the object. Finally, we try to transport some of the daily necessities using the prototyped robotic hand.

This paper discusses grasp stability of planar enveloping grasps with rolling contact. The grasp system is replaced with an elastic system. Potential energy of the grasp system is formulated. Grasp wrench and grasp stiffness matrix of the system are derived. The grasp stability is evaluated by the eigenvalues of the matrix. Curvature effect is analyzed by differentiating the matrix by the curvatures.

A robot gripper for correction work has been developed to automate rework of frozen foods and other food products. It is not known in advance which part of a food product needs to be reworked. Therefore, we developed a gripper that can rotate the gripped food in two directions in the hand so that the surface to be reworked can be found by rotating the gripped food in various directions.

Manipulation of flexible materials in wet conditions is one of the most difficult tasks in daily life and in factory automation. The technique of origami, in which water is applied to a flexible material such as paper or cloth to reduce unwanted wrinkles and create neat folds, is called wet folding. With this in mind, this study presents a robotic hand capable of wet folding operation. For the purpose, the gripper has not only conventional antipodal grasping function but also the functions of rolling and water supply. The efficacy of the developed system is validated through several experiences.

This paper describes a vacuum suction pad with a tactile sensor that can evaluate and predict the fall of a grasped object. First, the research background, usefulness of the suction pad, and related studies are described. Next, the principle of the conductive sponge substrate and amplifier substrate, which constitute the developed device, and the tactile suction pad fabricated by integrating them are described. Then, experiments to evaluate whether the developed device can ”judge whether an object is grasped” and ”predict whether an object will fall” are described, as well as the results and discussion of the experiments. Finally, the paper concludes and discusses future prospects.

This paper shows a new robot which can go thorough pipes. It is difficult to pass through long pipe lines fast. In order to solve this problem, we focus on Λ-drive, which are composed of a flat tube and a slider. We propose the robot which is equipped with some Λ-drives. This robot not only propels fast and long distance, but also supplies fluid which cleans pipes by carrying supply lines. We did some experiments that how long this robot can go ahead into pipes to decide the distance between Λ-drives.

In recent years, the improved response of hydraulic directional flow control valves and the miniaturization and weight reduction of hydraulic actuators have facilitated their application to robots, and research and development of hydraulically driven robots has been actively conducted. Hydraulic robots have large payload and ability to carry and tow heavy objects by taking advantage of the high F/M ratio of hydraulic-driven actuators. However, to handle hydraulics, various hydraulic components must be included, often making downsizing, weight reduction, and actuators a challenging issue. In this study, using the hydraulically driven robot "Tough Runner," which has been developed with the goal of independent long-distance walking, we conducted operation experiments in which control valves were integrated into the robot, and compared the results of the applicable range of the control valves based on the results.

We investigated whether stroking the skin near the external auditory meatus while listening to sounds enhances pleasant feelings. Previous studies have demonstrated that vibratory stimuli can affect emotions evoked by watching videos or listening to sounds. However, none have investigated the effects of stroking stimuli near the external auditory meatus, which is a location where the vagus nerves extend. Participants were exposed to three types of sounds, both with and without the stroking stimuli applied to their ears, and were asked to select the stimulus condition that induced the most intense feelings. When sound stimuli were presented together with the stroking stimuli to the ear, joyful and pleasant feelings were intensified, whereas depressing feelings weakened. Therefore, it was suggested that stroking stimulation of the outer ear in combination with sound stimulation can induce positive feelings.

In recent years, research and development using VirtualReality (VR) has been active due to improvements in video and audio technology, etc. By presenting tactile and force information to the user in VR, it is possible to improve the sense of presence, operability, and safety. In this study, we aim to develop a compact device that can perceive the impact position of a grasped object in a virtual environment by instantly changing the direction of the inertial force of motor rotation. In this paper, we compare the vibration waveforms of the hand when a ball is actually dropped onto a racket and when a tactile force is presented by the developed device, and conduct an experiment using a questionnaire.

In this study, we aim to efficiently intervene in psychological states with displaying devices that are highly interactive and capable of transmitting contact pressure distribution. In this paper, we report on a balloon actuator unit driven by a VCM and an elastic membrane to present contact pressure distribution stimuli to the palm. The experiment shows that the device has sufficient output and frequency characteristics to provide pleasant sensation. We have conducted a subject experiment to compare the balloon structures and contact states and find that the pressure, rather than force, may be acting on intensity perception. In the future, we investigate the effects of the pressure distribution stimuli using multiple units to improve the interactivity and variations of tactile information.

Human tactile perception is affected by individual differences in skin shape and material properties. In this study, we report the effects of individual differences in finger size on the strain energy density (SED) distribution in tactile receptors. We created seven finger models with varying widths and thicknesses using statistical data on Japanese finger dimensions. The deformation analysis of the fingers pressing against a rigid plate was conducted using finite element analysis software. The results show that, under the same displacement conditions, the SED distribution in the area between the epidermis and dermis tended to have a smaller maximum value with larger finger thickness. The width of the SED distribution tended to narrow with increasing thickness, and the length tended to lengthen with increasing contact force. However, further verification is needed because the effect of the shape of the finger pad is considered to be significant.

In this study, the spatial distribution of vibrations propagating from the fingertips to the wrist was obtained as a method of conveying the sensation of movement, such as a skill, and the effectiveness of this method was demonstrated by using multiple-point stimulation as a reproduction method. We have already confirmed that using the spatial distribution of vibration to discriminate between different contact movements and contact forces improves the results. The experience is expected to be improved by adjusting the intensity of signals at specific frequencies, such as noise reduction and target enhancement. This paper demonstrates the usefulness of using perceived intensity to adjust the intensity of stimuli with different frequencies.

Effective massage requires specialized knowledge and experience, and the number of massage therapists who possess these skills is limited. To solve this problem, massage robots with multi-fingered hands have been developed. Since this robot directly touches the human body, it is important to measure the force applied to the fingertips. However, there are few practical force sensors that are small enough to be attached to the fingertips of the robot hands and can withstand large loads such as massage movements. In this paper, we propose a vibration-type force sensor that is easily miniaturized and can withstand large loads such as massage, and discuss the prototype sensor and its evaluation results.

This paper reports on human skin detection using a heat-flow-type warmth sensor that discriminates the material of a contacted object by a heat-flow-peak value in a transient heat-flow waveform. In this study, first, a module composed of synchronous detection circuits was fabricated as a sensing circuit for high accurate sensing. Compared to a conventional voltage-dividing circuit, the variations of the heat-flow-peak values were reduced. Then, human skin detection was tested on hand and arm. The thermal conductivity estimated from the measured heat-flow-peak value was close to about 0.5 W/m･K, which is considered to be the thermal conductivity of human skin. The experimental results showed the possibility of human skin detection by the heat-flow-type warmth sensor.

We have been studying the application of fabric actuators, which are flexible and have a high degree of freedom of deformation, to tactile sensation presentation devices. In this paper, we focus on the friction of the fabric used for fabric actuators and report the relationship between the friction of the fabric in contact with the human body and the recalled sensation.

In this report, we describe fabrication and results of its preliminary experiments of a sensitivity-changeable flexible tactile sensor with magnetorheological fluid (MR) and electro-permanent magnets (EPMs). Many soft robots take advantage of pneumatic actuators to change their morphology. However, the hole system with pneumatic actuator tends to be large because of a bulky pump for the air supply. In addition, their response time larger. Thus, we focus on MR fluid whose viscosity can be changed in a short response time. A peristaltic pump actuates the MR fluid circuit. When the EPMs are activated, MR fluid’s viscosity increases, which induce the tactile sensor’s morphological change. We conducted some preliminary experiments and confirmed that the fabricated system can be used for the sensitivity change of the tactile sensor.

CycleGAN is a technique that realizes image transformation by learning the relationship between the domains of two images. CycleGAN is good at style conversion such as color and pattern, but there is a problem that conversion accompanied by shape conversion is difficult. The reason is that CycleGAN recognizes the image background as part of the conversion target and cannot perform feature extraction well. In this research, shape transformation by CycleGAN is performed by preparing a data set in which the image background is deleted by a generative model.

In recent years, automated driving has been attracting attention, and various companies and institutions have been actively conducting research on this technology. In particular, automatic driving requires high-precision location estimation, which requires a 3D point cloud map with high-precision information on the road surroundings. SLAM, which simultaneously estimates self-position and maps the environment, has been attracting attention as a method for constructing 3D point cloud maps, but it is known that 3D point cloud maps constructed by SLAM have residual position errors. However, it is known that 3D point cloud maps constructed by SLAM and other methods have a problem of residual positional errors. Therefore, in this study, we aimed to detect degraded locations in 3D point cloud maps by focusing on errors in the translational direction of point clouds, taking advantage of point cloud clutter.

In this study, purpose is pose estimation of flexible objects. The object is a cord-like flexible object, and moving image analysis using deep learning is used. The method proposed is markerless pose estimation. The movement of flexible objects is captured as a video and labeled on the image. It learns from labeled images and estimates labeling points. A curve is calculated from that point, and the pose is estimated. In this paper, we actually photographed the movement of a flexible cord-like object, and compared the estimation result with the actual photograph for verification. It was shown that a posture similar to the actual posture can be estimated. The maximum tip error was 2 cm.

Visual Teach & Repeat Navigation (VT&R) is a traditional way to let a mobile robot navigate without explicit localization. Most VT&R methods perform image matching between images, which the robot obtained while operated by a human (teaching process) and while running by itself (repeating process), in various ways including classical image processing and deep learning. In this paper, we investigate how the image matching methods work when considering outdoor VT&R scenarios with extreme appearance changes of environments. Our offline experimental results suggest that deep feature point extraction and matching techniques such as SuperPoint [1] and SuperGlue [2] help accurate repeat navigation even if illumination condition drastically changes from teaching process.

In the field of product development, shortening of process time is desired and simulation such as CAE is essential. However, the product shape modeled by CAE is an ideal shape without irregularities and is different from actual products. Therefore, generation method of surface models with irregularities (skin model shapes) has been developed using DTCWT (dual-tree complex wavelet transform). Skin model shapes are models represented by points, polygons, or functions, and often used in tolerance analysis. In this study, we compare skin model shapes generated on a computer with the surface of the actual processed material and verify the validity of the method.

In this research, we aimed to develop a measurement method using a laser as a method of three-dimensional measurement. The object was irradiated with a gradient laser and the situation was photographed with a camera. The changes in the laser pattern were digitized by the hue obtained from the camera image, and the shape of the object was measured in three dimensions based on the amount of change.

This study analyzes the recognition process of examining the texture of fine particles through tactile sensing using human fingers and aims at establishing the structure of texture recognition. Our previous study confirmed the occurrence of hysteresis in tactile sensation in the mid-teens. The research study describes the influence of hysteresis in tactile sensation on KANSEI evaluation of fine particles.

Pseudo Haptics is a method of presenting tactile sensations without using any haptic device. However, Pseudo Haptics is problematic because of its limited effect size and significant differences in effect among individuals. To improve them, we investigated the effect of perspective increase using the line perspective illusion that affects Psuedo Haptics. For this purpose, we examined the perceived distance to the object in VR space using the linear perspective illusion and then the perceived weight and movement distance when the object was pushed. The results showed that the illusion could not significantly change the perceived distance and that it was difficult to improve the weight perception by Pseudo Haptics. This may be because the observers could not consciously perceive the illusion, and it was difficult to notice the illusion due to the small change in the visual object caused by the line perspective method.

Soft robots enable various functions with their simple structure. However, typical flow control using many solenoid valves increases the complexity, mechanical structure, and cost of the control system. To address this issue, in our previous study we developed a device reversing the flow direction, switching blowing and suction with a single air flow control. In addition, we found the device can be used as force and strain sensor. In this study, we investigated the sensing functions.

In order to overcome the aging society of Japan, the robot for daily tasks are widely researched and developed. In particular, the pick-and-place is the popular tasks well researched since the dawn of robotics, such as sorting of bulk goods or packing of prepared foods into lunch boxes. In recent years, soft robotic hands with tactile sensing have been attracting attention for grasping a variety of objects. As one of those hand, Universal Vacuum Hand (here in after UVH) utilizes the volume resistivity of the conductive powder. UVH consists of the elastic membrane bag with the conductive powder, that can detect the contact with the object to be grasped, but it is difficult to measure applied force to the object to be grasped. In this paper, we propose a novel measurement method with mixed powder of conductive and non-conductive ones. To evaluate the load applied to the UVH from the volume resistivity of the mixed powder, we report the effects of measurement elements, such as vacuum pressure, shape of electrodes.

In this paper, we focus on vibrotactile stimuli for the purpose of changing and amplifying emotions, and evaluate the changes in emotions caused by vibrotactile stimuli converted from music. To construct a system that simultaneously presents music and vibration, we developed a chair-type vibrotactile display and an automatic conversion method of vibrotactile stimuli based on music. The method generates stimuli suitable for tactile perception by shifting the pitch and dividing the music. In an evaluation experiment, we evaluated the changes in the emotions elicited by the music when vibrotactile stimuli were applied using temporal and posterior evaluation methods. The results showed that not only were the emotions strongly caused by the music, but they were also amplified and changed by the presentation of the vibrotactile stimuli.

This study investigates the effects of the facial expressions of an avatar on haptic perception during human-agent interactions by means of physical and psychological measures. Thus far, we demonstrated that changes in facial expressions can generate haptic perception in a system in which a human-like agent on the display behaved as if it pulled the mouse cursor against a human participant control. This paper presents an experiment to evaluate the emotions recognized by the facial expressions designed by the authors. This would enable us to study the relation between recognized emotions and haptic perception.

This study focuses on the importance of the tracing sensation, which helps in identifying surface texture and frictional cues. However, the traditional methods to render this sensation through real tracing movement using a plate or roller are often limited by the size of the devices. The study presents a new approach where a fingertip-sized disk is rotated while in contact with the finger to create a tracing sensation. Results indicate that this method effectively provides a realistic tracing sensation continuously and in any direction. The study also investigates the impact of reversal movements in the direction of rotation on the realism of the tracing sensation, finding that such movements may be particularly effective for tracing rough textures.

Numerous studies have explored the body tilt and sway elicited by vibratory stimuli, which are thought to be related to kinesthetic illusions. However, prior studies have not thoroughly explored the conditions that change the interpretation of self and environmental factors. In this study, subjective body sway was induced through alternating vibrations that were applied to the antagonist ankle muscles as the switching frequency of the vibrations was changed. Results indicated that an illusion induced by a low switching frequency tended to be interpreted as self-sway, while a high switching frequency led to the interpretation of environmental sway, which suggests that when individuals receive sensory information that cannot be realized by their common movements, they interpret it as environmental sway through the process of elimination.

We have focused on mecanum wheels among various omni-directional wheels and conducted load capacity experiments with commercially available mecanum wheels. As a result, it was found that the wheels were easily damaged by the change in propulsive force due to rotational friction. Therefore, we proposed two omni-directional wheels with different types of roller patterns that improve load capacity by increasing the roller tip diameter without changing the wheel diameter. Experiments were conducted on four types of wheels: the two proposed wheels, a commercial wheel and a wheel with a normal roller pattern, using a three-wheel experimental vehicle. Current values and robot trajectories were evaluated for their loading capacity, and we confirmed that the proposed roller patterns effectively improved the load capacity of the Mecanum wheel.

In this research, a leg mechanism of a micro hexapod robot using an origami structure equipped with electrostatic angle sensors were fabricated and evaluated. The walking mechanism has two degrees of freedom, the vertical and horizontal movements of the leg. The reliable walking motion was achieved due to the angle feedback directly.

In order for robots to play an active role in human daily life, robots are required to interact with humans. Manipulators are required to be safe in case of contact while making quick movements. In-link actuators are useful as direct-drive actuators with low inertia. In addition, we developed the joint actuator that is easy to disassemble and assemble by taking advantage of the feature of the in-link actuator in which stator and rotor are not treated as a single motor, but are separated into each link. Based on the design intent of the in-link actuator, the aim is to create a joint actuator that simultaneously satisfies the structural advantages of low inertia, light weight, small size, high back-drivability, and reduced number of parts, as well as ease of disassembly and assembly.

Efficient transportation of objects by robots is required due to the declining working-age population. This study focuses on sphere transport as a means of efficient object transport by robots. To date, there is no methodology for mechanism design based on kinematics properties for sphere transport by robots. Therefore, we propose a method for designing a sphere transport mechanism based on kinematic properties. The paper presents the requirements and methodology for designing a sphere transport mechanism based on kinematic properties.

In recent years, the development of multifunctional mobile platforms has gained momentum with the aim of reducing transportation consumption. In the previous study, a mobile robot named UDOn (Ubiquitous Delivery On-demand robot), with the ability to go up and down stairs, was developed. To further enhance the transportation capacity of the mobile robot, this study developed a robotic arm mechanism for UDOn, allowing it to autonomously grab goods. To achieve the concept of reducing development costs, the robotic arm utilizes only four degrees of freedom and is structured to maximize the use of each degree of freedom for completing the required tasks. The presented method and experimental results in this paper demonstrate the feasibility of such a robotic arm.

Injuries during the practice of sports in gymnasiums have been reported, and one of the causes of injuries is due to environmental factors as splinters on the gymnasium floor. Therefore, it is important to regularly inspect gymnasium floors. However, it is difficult for humans to visually inspect the entire gymnasium floor. Thus, we have developed an automatic inspection system. In this study, YOLO was proposed as the detection method for markers attached to splinters. The detection ratio of the proposed method was improved by 25.0 % compared to the conventional method of threshold color segmentation process. In an inspection of an entire gymnasium, the proposed method detected 33 markers was able to detect splinters that could cause injury.

In this paper, we further develop the concept of a reflexive drive robot mechanism that we have been working on for some time, and propose and realize a complete mechanization of a contact counting element that is able to extract the mechanism of the leaf closure initiation and sea cucumber characteristic switching of the flycatcher, as well as the flexible-rigid switching mechanism. Through experiments using the actual machine, we confirmed the basic validity of the proposed principle, including the contact counting function and the reflexive drive.

Autonomous detection for gas leaks in chemical plants is significantly important. In this work, we address this issue by using multiples robots: ground robots and drones, which have complementary advantages and disadvantages in gas sensing. Then, we develop a stepwise search system from plane to height, taking into account the operating time of the robot and its impact on airflow. In the first step, we use a ground robot to perform two-dimensional location for gas source on the ground, which is followed by a height search strategy for gas sources by a drone as the second step. The proposed system was also verified in an actual indoor environment for the case of an ethanol gas source fixed at a height to locate the gas source.

In this study, we developed an outdoor surveillance robot system using public 5G. The robot patrolling outdoors is connected to a remote observer via 5G and ROS2, and the observer control the robot while monitoring camera images remotely. We tested three network configurations to realize remote communication. Experiments for remote autonomous surveillance were conducted.

This paper proposes a distance estimation method for a gesture interface with a monocular fovea-lens camera to control AMR position. The interface recognized gestures based on the joint coordinates obtained from the skeleton estimation library. From the acquired data such as the angle and length of the arms, the interface classified 10 ground coordinates centered on operator. This research developed an image-based distance estimation method that used only a camera as a sensor. This paper presents an estimation method that considers the special curvature of the fovea-camera.