This paper presents an object arrangement model and grasp strategy for object picking. Daily life objects are usually surrounded by other objects. The nearby objects around a target are obstacles to object picking. Therefore, differing grasp strategies are needed depending on the object arrangement. In this paper, we classify object arrangement pattern based on which target surfaces can be accessed by a finger. And we discuss applicable grasp strategies for picking up the arranged object.

View full abstract

Secure grip by conventional robot hand without any damages for the soft and easily damaged object such as strawberry is very difficult to realize. This study aims to develop a novel robot hand (finger) for the grasping of such soft and indeterminate form objects. As the methods for grasping such soft objects, soft material covered finger, increasing contact area by multiple joints mechanism, sensing and control of the contact force are used and developed. However, it is not so easy to realize a stable grip for indeterminate form objects. Most existing robot hands do not have sufficient performance for the soft and easily damaged object on the stress of contact surface. This article proposes a passively deformable soft finger to grip such easily damaged and indeterminate form objects. The hand uses a soft bag in which a fluent material is filled. This hand enables uniform contact pressure distribution for the contact surface by deforming the shape of the hand passively and by hydrostatic pressure inside the bag. The contact pressure distribution is measured to evaluate the performance of the hand.

View full abstract

In order to detect a slip, various slip sensors have been proposed so far. However, they are not in practical use yet. There are two causes. One is the difficulty to mount many sensors in downsized robot hands. The other is the difficulty to maintain the many sensors. We propose a slip detection method with the multi-axis force/torque sensor in order to resolve these problems. The multi-axis force/torque sensors are mounted in general robot fingertip and the above causes are cleared.

View full abstract

In the study, we propose high positional precision serial link type manipulator which use singular configuration properties that endeffector doesn't move near singular configuration even if joint angle changes greatly. In this paper, we describe that instance designing of 1 DOF mechanism which can change number of singular configuration and position of end effector of singular configuration on mechanistic parameter.

View full abstract

We developed Transformable Robot Hand that the fingers can be attached / detached and changed placement of finger-base freely. In this paper, we present the system of Transformable Robot Hand, that is installed a camera recognizing shape of an object, and how Transformable Robot Hand transform placement of finger-base to the most suitable to grasp an object.

View full abstract

In recent years, the robot has been in the field of medical scene and human life. These robots operate in an environment close to human, it is necessary to achieve the coordination of work and ensure the safety of people, when these robots contact with humans. Therefore, these robots be flexible and lightweight, it is desirable to have the characteristics a close to human behavior. For this reason, we have studied and developed an artificial muscle actuator using straight-fiber-type artificial muscles derived from the McKibben-type muscles. In this paper, we propose a new robot hand by artificial muscle actuator that is thin straight type like a wire.

View full abstract

We manufacture an anthropomorphic hand with five fingers and control the fingers consisted of the multi-joints to grasp various objects by wire driven actuator. The state of grasping and holding objects can be detected using servomotor torque. There servo's torque was measured using the current sensor on the servomotor. In this paper we will observe the action and the response of each servomotor based on the current sensor data. We derived calibration method for current sensor. Additional, we did the tensile test of Kevlar yarn. Also, calculate the torque of servomotor used by current sensor data. Furthermore we show the mechanism of the designed robot hand.

View full abstract

This paper is about a teaching system for multi-fingered robot hands using kinetic information of a manipulated object with feedback control. The purpose of this system is to operate multi-fingered robot hands with high DOF intuitively and effectively. The system calculates the appropriate trajectory according to the kinetic information of the manipulated object under the twist rolling contact constraint. For more stable manipulation, visual and force feedback control using kinetic information of the real manipulated object is appended to this teaching system. This paper describes the method of feedback control and the future works are discussed.

View full abstract

The development of the robot industry has a remarkable. In the field of various industries, a robot working in place of human is expected. Therefore, we have developed a small and five-fingered robot hand like a human hand. Then, we built the master-slave teleoperation system in order to work various tasks. However, there is problem in simple teleoperation systems to do the pinching task with stability. The problem is not to have the sense that the operator actually grasps. It makes instability of grip force and the problem that the pinching task cannot be done with stability. In this paper, we propose a teleoperation system that used the grip force stabilization method. Here, the grip force stabilization method is used for moving in the direction where the change in the external force is cancelled. We show the effectiveness of the suggestion method by experiment.

View full abstract

Actuation and control of robotic hands is one of the difficult problems in mechatronics that still needs to be solved. In this paper we present a kinematic analysis of a robotic finger that is actuated by twisted strings mechanism that we named Twist Drive. The principle of actuation is briefly described in the paper and a structure of the developed robotic finger is presented. Forward kinematics for a finger with nonlinearly coupled joints is given. The obtained Jacobian is used in force control on the finger's tip. Experimental results are presented. Force control in two directions was successfully demonstrated.

View full abstract

In this paper, we propose a simple structural constant torque spring using cross-sectional deformation. We call this Cross-Sectional Deformation Spring (CSDS). The CSDS consists of cutting a part of the elastic tube. In this paper, we experimentally confirm how the parameter that composes CSDS influences the characteristic. For these result, the CSDS has possibility to achieve constant torque characteristic. And we introduce implementation to the finger part of the robot hand as an applied example of this CSDS.

View full abstract

The objective of this research is to detect posture of an object using the Net Structure Proximity Sensor applied to the fingertip of robot hand for reliable grasping. In object grasping and object sensing, occlusion occurs when using the vision sensor around several millimeters from grasping object. The solution of this problem is to introduce the proximity sensor which detects the object position and distance in near range area. The posture detection uses the sensor position output which takes zero not being affected by the distance whenever the surface of object and the surface of the sensor are parallel position. This detection technique is introduced to manage the object grasping successfully in adjusting the robot hand posture in two axes to appropriate grasping position.

View full abstract

This paper presents an experimental verification of the externally sensorless dynamic object manipulation which has been proposed by us. This method enables robot to achieve stable object grasping and manipulation in dynamical sense without any external sensing. However, the effectiveness of the proposed method has been shown through only numerical simulations, so far. In this paper, a practical usefulness of the method is shown by an object manipulation experiment using a prototype. The prototype developed in this study has three fingers and each finger has 4 DOFs. By using this prototype, an experiment for stable object grasping with position and attitude control without any external sensing is conducted. Through this experimental result, the usefulness of our proposed method is demonstrated.

View full abstract

This paper describes an oblique feed screw which can be used as load-sensitive continuously variable transmission (CVT). This CVT consists of a screw, spring, and bearing, and is remarkably simple and compact. Its reduction ratio change automatically in response to the load. We have developed the CVT of 13.8 [g]. We have experimentally verified that it can exert a very strong force of more than 100 [N], and the CVT can increase its reduction ratio form 20 to 45.

View full abstract

In the industry, automation has been progressed by robot that can work more accurately and quickly than a human. However, most of them are limited to simple work. On the other hand, human being can handle various tasks with a variety of tools. If robot can handle various tools, wide range of tasks can be handled. In our previous works, the robot hand for nipper type tools has been developed. This hand is moved by only two motors, but it can handle various nipper type tools. In this paper, several experiments are executed to confirm its feasibility. In addition, an approach to estimate joint position is discussed.

View full abstract

This paper proposes effective utilization of contact information for robot manipulations. Manipulation sometimes needs not only contact force control but also control of contact force direction. In particular, grasping a spherical object by fingers contacting on the points of upper hemisphere using friction force needs to consider direction of contact force. This task needs to consider slipping at the contact point and magnitude of force to resist gravity force. This paper describes a theory about how to grasp a spherical object using friction force by two fingers on the surface of upper hemisphere focusing on angle of contact surface and contact force direction. This paper also shows our robot actually grasps a spherical object by two-finger manipulation, and the result of magnitude and direction of force at experiments.

View full abstract

This paper describes the finger mechanism with omnidirectional driving roller to realize the two active rotational axes on the surface of the grasped object. As the omnidirectional driving roller, we adopt the Omni-Crawler we developed: the crawler mechanism with circular cross section. This cylinderical tracked unit can be used as the roller and the fingers with this roller can manipulate the grasped object in the arbitrary axes. The basic concept of this finger is proposed.

View full abstract

This paper describes the morphing omnidirectional gripper which is able to grasp various objects with low melting point alloy, functional fluid, dilatancy fluid, and so on. The deformable part of the gripper changes its shape by covering all direction of objects and makes the contacting area higher. This time, we especially focus on the "Hot Ice" phenomena to realize higher grasping motion. The basic performance of fluid of CH3COONa has been observed.

View full abstract

In previous researches, it is difficult for developed grippers to grasp spherical objects and grasp object after grasping continuity. This paper discusses omnidirectional enfolding gripper mechanism to realize above two functions. At first, we examine basic performance of this developed gripper through the experiment of the single object grasping. Next, we realize the composition of the basic motion of this gripper to grip plural spherical object continuity based on the results of the previous single object grasping experiment. In addition, we realize that this gripper can grip various geometric object and softness objects continuity using a same motion.

View full abstract

In Japan, the penetration of electric prosthetic hands is low, that is around 1%, and the hand is not popular. Authors have developed a myoelectric prosthetic hand which has high grasp force and 2-DOF wrist based on a survey of demands for prosthetic hands. This paper presents a novel myoelectric prosthetic hand and an artificial skin. The prosthetic hand is improved from the previous one. Experiments are performed to show characteristics of the prosthetic hand.

View full abstract

Sensing of joint torque is necessary for dexterous grasping and manipulation with a multi-fingered robot hand. Some methods for measuring joint torque are proposed. Authors have proposed a small torque sensor. The torque sensor can measure strain of the elastic body. It can be built in the robot finger mechanism. Torque measurement system requires the miniaturization of both sensor and amplifier circuit. This paper reports a new torque sensor system which has a built-in amplifier circuit. The experiments are performed to show characteristics which are linearity, hysteresis and frequency response. Motor can be controlled using a feedback signal of the torque sensor.

View full abstract

In this paper, we studied 2D caging manipulation by robots and walls in a real environment. Manipulation of a circular object was successfully performed by two or three circular mobile robots controlled with a camera. However, caging manipulation failed in some cases due to jamming. This problem can be solved partially by reducing friction, but further investigation is necessary.

View full abstract

This paper addresses the adaptive navigation problem for a swarm of autonomous mobile robots. From the observation of schooling behaviors, we propose a novel adaptive navigation algorithm, enabling large-scale robot swarms with limited sensing capabilities to navigate toward the goal. The adaptive navigation algorithm is based on our geometric local interaction which allows three neighboring robot to locally form an equilateral triangle lattice. During navigation in an unknown environment, mobile robots maintain the local formation. According to environmental conditions, this algorithm allows the swarm of robots to split into multiple groups or merge with other groups. More specifically, an encountered dead-end passageway drives robots to travel in the reverse direction, and they merge with another adjacent group. Accordingly, their local motions can recover the whole swarm network from the constrained environment. Through extensive simulations, we validate the effectiveness and scalability of the proposed algorithm.

View full abstract

A systematic design methodology of swarm robots that exploit physical interaction is still lacking. To tackle this problem, we propose a design method of swarm robots that exploit hydrodynamical interaction. The robots are placed on a 2D plane filled with compressive fluid, and they propel toward an attractant by pushing their bodies against fluid. Local sensory feedback is implemented into each robot, where its moving direction and driving force is adjusted based on the pressure gradient of fluid. Owing to this sensory feedback mechanism, the robots can swarm orderly without any informational interaction. The validity of the proposed design method is investigated through simulations, and the results show that robots successfully swarm even when initial conditions and direction of an attractant are varied.

View full abstract

This paper presents a fluid-filled modular robot inspired by a living coupled oscillator system constructed by plasmodium of true slime mold. Each module of the robot has a truly soft and deformable body, stemming from real-time tunable springs, filled with fluid. The robot consists of these homogeneous modules which are physically coupled with fluid-filled tubes. Exploiting this physical long distant interaction, the robot is capable to induce versatile oscillatory patterns and transitions between the several oscillatory patterns in a fully decentralized manner. The simulation results obtained are expected to shed new light on design scheme of life-like robots to reproduce astoundingly versatile, supple and robust behaviors.

View full abstract

In exhibition space, it is possible to control human flow implicitly by changing the layout of exhibits. The objectives of this paper are optimization of exhibit layout to reduce the congestion and amenity space design. For these purposes, macro modeling of human swarm behavior and optimization method of exhibit layout are required. So far, human swarm behavior has been modeled by two-dimensional vector field and individual behavior is represented by dynamics including collision avoidance vector of individuals. In this paper, we extend the human model to multi-dimensional dynamics to represent visitors' stopping to view exhibits, and the layout of exhibits is optimized based on the proposed model by minimizing the collision avoidance vector. The proposed method is verified by simulation and experiment using swarm robots which consist of autonomous mobile robots and radio controlled robots.

View full abstract

In the densely-populated urban areas, pedestrian flows often cross each other and congestion occurs. In order to reduce the congestion or the risk of an accident, it is required to control swarm behavior of pedestrian. This paper proposes an implicit control method of the crossing pedestrian flows by moving guides. Each pedestrian is modeled with a particle, and virtual density and temporal/spatial frequency are computed from position of particles. From analysis of relationship between guide motion and temporal/spatial frequency, we verify the validity of control method, which is originally introduced from analysis result with the continuum model.

View full abstract

In this paper, a motion-planning scheme that enables robot arms to avoid structural damage is described. By using this scheme, the robot arms are encouraged to protect themselves from structural damage by searching for a safer attitude when their structural risk becomes high during their given tasks. The structural risk is determined by using two parameters, i.e., yield function value and total strain energy in the architecture. Furthermore, more strict risk judge is enabled by considering inertia in target motion calculation algorithm.

View full abstract

This paper presents a new index that allows us to evaluate the manipulating force accuracy of manipulators considering the task-directions. First, the new index (FAIT: Force Accuracy Index for Task-directions) is derived. Second, the geometrical relationship of the FAIT to the manipulating force ellipsoid is represented. Third, some postures of a 3-DOF planar manipulator for a task are evaluated on the basis of condition number and the FAIT, respectively. Finally, form these evaluation results, it is shown that the FAIT is useful to evaluate the force accuracy of manipulators for the task-directions.

View full abstract

We propose an algorithm for automatic motion generation of a manipulator using virtual repulsive forces whose size is defined on the basis of the distance between the manipulator and obstacles, and its force direction is defined as outward way from the surface of obstacles. The motion path of the end effecter is modeled as the structure with mass points and springs, and its form is changed by the repulsive forces to generate collision-free path. The collision-free posture of the manipulator is also generated using repulsive forces, the posture is modified by the amount of joints torques which is transformed from acting repulsive forces. We have applied this algorithm to the 7-DOF manipulator, and it is confirmed that this method is able to automatically generate collision-free approach and trace motions.

View full abstract

Manipulator system can improve the working efficiency in a factory. An operator should teach points of path for a manipulator when he starts up it. A teaching pendant is generally used to teach points for a manipulator. However, the teaching pendant has required specified skills for the operator and is not intuitive. Therefore we develop an operation device to control a manipulator with human hand motion. An operator can control a manipulator intuitively with proposed device. However, the manipulator sometimes moved wrongly. One reason is that angle around vertical axis rotation is not correct. So, we invent a new operation device with a camera module and propose fusion algorithm of image information and inertial sensor to correct the angle. The effectiveness of proposed algorithm is confirmed through experiments.

View full abstract

This paper addresses the problem of shapeless grasping for a home-use logistical system where a shapeless stable caging approach is forwarded to control ceiling robotic hands. We demonstrate that stable caging is essentially an immobilization grasp that employs least number of fingers by introducing CC obstacles and perform stable caging with least finger number decided by inner Largest Empty Circles (inner LECs). Our system works well with object boundary data generated from one laser scanner and demonstrates its robustness in experiments on daily utilities.

View full abstract

This paper investigated the trajectory generative capacity of the kinematics model by all integer variable programming and discretization. We compared the computational results of the general inverse kinematics model to the integer inverse kinematics model, and their results confirmed the similar both outputs and simplify the process.

View full abstract

Pulse-echo method is widely used as a ranging method with ultrasonic wave sensor. But faint reflected wave tends not to be perceived in this method. Then, we have proposed integration-type ultrasonic wave sensor to use faint reflected wave effectively. This paper proposes a method of path generation by using the sonar ring which consists of some integration-type ultrasonic wave sensors. This sonar ring can shorten sensing time because it can obtain information from plural directions simultaneously. When the sensors transmit ultrasonic wave at the same time, interference between sensors may occur. However, experiments showed that sub-goal to avoid obstacles was generated by using integration value even if ultrasonic waves were transmitted at the same time.

View full abstract

We have developed a path planning and path following system for a mobile robot. First, the point cloud data of the environment are transfered into polygonal expression by giving the vertex using GUI. Considerating the robot size, the planning system calculates the approximated polygons of configuration obstacles in the two dimensional configuration space, and generates the visibility graph and tangential graph. Then, the shortest path between given start and goal point is planned by graph search. We also developped a control program for our experimental mobile robot to follow the planned path. We made experiments of path planning and travelling control in our laboratory.

View full abstract

Liquid in container slops around by change of acceleration, so robot can't transfer liquid container as the rigid box. We suggest control method to transfer liquid container with liquid slop inhibit. Control inputs are horizontal and angular accelerations of container. When the ratio of water depth to container width is small, liquid flow is expressed nonlinear equation of state model. We applied Receding Horizon control to it. And we proposed a new method of trajectory generation to inhibit liquid slop. It is applicable to both deep and shallow case because this method doesn't use fluid model. We examine liquid motion's change caused by changing center of rotation and find a optimum it. We verified the efficacy of control by simulating with MPS method that was used incompressible fluid analysis.

View full abstract

We propose a fast motion controller for a robot which has a flexible arm using the IDCS control scheme. The IDCS control scheme, which corresponds to a numerical computation of the inverse dynamics of the systems, allows stable high performance control even when the system model is known with inaccuracies. We applied in simulation the IDCS controller to the robot arm and compared the performances of the IDCS with PID controller. We can get good simulation model of the robot for the IDCS by estimating inertial parameters using generalized coordinates of the baselink, the joint angles and the external forces information.

View full abstract

This paper presents a method to cancel the mechanical sway which results from the reaction moment of rotary motion. The torque cancelling system (TCS), which consists of a motor and a rotating body, generates the required moment to remove the sway which is calculated by the parallel solution scheme. The system needs to supply more torque and be small and light enough when installed in a practical robot. We evaluated the effect of the TCS with a geared motor and confirmed that the feedback control is necessary to compensate the backlash in the gear.

View full abstract

The conventional demand on robots in industrial fields has been oriented mainly to mass production, but the demand nowadays is diverse to meet the trend toward high-variation and low-quantity production. To meet this demand, we have proposed a reconfigurable robot which can change a robot configuration easily. In this paper, we construct the multirate sampling control for a reconfigurable robot. Firstly, we describe the concept of a reconfigurable robot. Secondly, the decentralized control in the system consisting of local and central intelligence is explained. Thirdly, the advantage of the multirate sampling is explained. Lastly, the effectiveness of the proposed system is confirmed by several experimental results.

View full abstract

This paper presents an algorithm for the control of the position and the orientation of the end-effector of a rigid-link manipulator. The control algorithm is an extension of PID control, and ensures smooth, exponential resuming motion from large positional/orientational errors that result in the saturation in the actuator torques.

View full abstract

In this paper, we propose synchronized hopping control method of human and robot by using forced entrainment of nonlinear oscillator. First, we show that periodic input control system is self-excitation system same as a van del Pol's equation. Next we describe synchronized method between human and robot by using forced entrainment. Finally, we show results that verified effectiveness of our method by real robot.

View full abstract

Robotic systems are practically subject to some physical constraints. In particular, violation of joint limits can deteriorate control performance and lead to breakdown of the system. We have proposed the maximal admission of joint range of motion using redundant degrees-of-freedom for kinematically redundant manipulators. This paper presents its application to image-based visual servoing. The controlled object is an eye-in-hand system, which is composed of a three degrees-of-freedom planar manipulator and a single camera that is mounted on the manipulator's end-effector. The effectiveness of the application is evaluated experimentally.

View full abstract

This paper describes intersection recognition in urban environments to acquire appropriate motion decision for autonomous navigation. Intersection shape which includes width, curvature, position and heading of a road, is given by road geometry scanned from onboard laser sensor. It is proposed that an intersection existence can be classified from constructed map. The proposed method is implemented and the effectiveness of the intersection recognition is verified in outdoor environment.

View full abstract

This paper addresses optimal motion planning for a machine. We discuss a problem of discretization error for the planning in continuous state space. We propose a method to divide state space considering state transition according to a mechanical model. In the proposed method, we define discretization error as a divergence of state transition and we organize an optimal discretization that minimizes discretization error using a binary tree. To validate the proposed dividing method, we conducted several motion planning simulations in phase space using double inverted pendulums. As a result, the planned motion in discretized phase space by the proposed method showed higher optimality than that by uniform discretization on equal calculation time.

View full abstract

This research aims to construct assembly planner for complex shaped objects. For the first step of this research, this paper explains the method to construct the contact state graph. By clustering the polygon model of the assembled object, the surface of the object is split into several regions. Then, by simulating the assembly sequence, we consider recording in which region of clusters the contact point is included. The contact state graph is constructed based on the information of the cluster where the contact point is included.

View full abstract

In this paper, we deal with the fault-tolerant control system for a reconfigurable robot. Firstly, the concept of a reconfigurable robot is described. Secondly, the decentralized control system consisting of local and central intelligence is explained. Then, to improve reliability of a reconfigurable robot, we propose the control system to detect a fixed fault of the robot and to maintain control performance under the failure. The proposed control system is realized by using a fault detector constructed by a disturbance observer and simple inverse kinematics. Finally, the effectiveness of the proposed system is confirmed by some experimental results.

View full abstract

This paper proposes how to acquire successive jumping on artificial creature like a locust under the physical virtual environment. A concept of "behavior simple (BS) and behavior composed (BC)" is applied to the model. In order to realize successive jumping, its behavior is divided into two BSs, taking off and recovery motion. After two BSs are trained by use of the evolutionary artificial neural network (EANN), successive jumping can be acquired as BC.

View full abstract

Swarm robotic systems (SRS) are a type of multi-robot systems, that consist of many homogeneous autonomous robots without any global controllers. Since a task given to this system cannot be achieved by a single robot, emergent cooperative behavior is expected in a robot swarm by a certain mechanism through the interactions among the robots or with an environment. When robots develop meaningful collective behavior, it is very important to analyze autonomous specialization in a SRS. In this paper, we propose a novel method for analyzing the robot group behavior for SRS based on clustering the complex network associated with robots in the cooperative package pushing problem. The use of this method allowed us to visualize of autonomous specialization and analyze the robot group behavior of SRS.

View full abstract

Reinforcement learning is effective in acquisition of optimal control policy. However, the calculation amount increases in high-dimensional space. In this paper, we propose a global and local optimal control method using dynamic programming(DP) and differential dynamic programming(DDP). In the global part, approximate the optimal trajectory in the state space by DP. In the local part, optimize the approximate trajectory in the neighborhood by DDP. The proposed method can reduce the calculation amount in optimal control.

View full abstract

The dynamic plane covering by swarm robots is a problem to make a trajectory of move multiple robots in a given field so that the trajectory covers the field, which often appears in agricultural works such as water spraying. We solve the problem by applying the swarm robot concepts in the way that a single robot called a leader moves around in the filed and all the other robots called followers try to catch up the leader. The advantage of this solution is that we only need to specify a trajectory of the leader and we expect the scalability of the number of the robots. On the other hand, we have to carefully design the trajectory of the leader and the connectivity network between the followers so that all the robots are always connected and under control. This paper presents the relation between the connectivity network, the control stability, and the covered field both analytically and numerically.

View full abstract