This paper describes a retraction control scheme for a space manipulator after grasping a floating object in space. Many control methods for a pre-retracting phase have been proposed by other authors. However those for a retracting phase have been very few, and the essential problem which an object is moving had not been treated so far. Firstly we divide a retracting process into four phases, and clearly show the problems to be resolved. Secondly we derive a new kinematic equation which relates angular velocities of joints with the integral of force and moment at an end-effector in space. Thirdly a force control method is proposed by using that important equation, and the possibility to grasp within the admissible error is proved effective: This control method corresponds to a velocity control for an end-effector feedbacking the integral of force and moment considering the derived kinematic equation. Fourthly we apply that control method to a retraction control, and demonstrate the stability of the proposed control scheme. Finally the effectiveness of the proposed retraction control method is shown by computer simulations.
Moving an object on a work surface by pushing it is one of the operations which are required in the robotic manipulation. The motion of the pushed object is closely related to the distribution of the frictional force between the object and the surface. When the pushing force vector applied to an object passes through the center of friction which is the centroid of the friction distribution, the motion of the object is a pure translation. In this paper, we describe a method for estimating the friction distribution of an object and the position of the center of friction by pushing the object several times using a manipulator. First, on the assumption that the object has several points in contact with the work surface, we suppose the position of the points. By this supposition, we derive a linear equation relating the pushing force and the friction distribution. Then, substituting the measurements obtained from pushing the object several times by the manipulator into this equation, we estimate the friction distribution. Simulation and experimental results show the effectiveness of the approach.
An efficient computational algorithm for motion control of multi-arm free flying robots in space is discussed. For the motion control of the space manipulator mounted on a spacecraft, the movement of spacecraft itself must be considered. The generalized. Jacobian matrix is one of the good ways to solve this problem. Though an efficient computational algorithm of the generalized Jacobian was shown, it does not have a good prospect to extend to the case of multiple arms. For dexterous manipulating tasks, cooperative manipulation by multi-arm manipulators is desirable. In this paper, we show that the generalized Jacobian for multi-arm space manipulators can be obtained efficiently by regarding the total multiple-link system as a composite system which consists of two links with one joint. The proposed algorithm is applicable to any tree-structured multiple arm manipulators with not only rotational joint but also prismatic joint except close loop structure. We also show an efficient computational algorithms of the resolved acceleration control for multi-arm space manipulators.
Postural control of a falling cat was investigated analytically and experimentally. The authors intended to develop such a robot as a cat that could restore her attitude at free fall by twisting and controlling the body. There are still contradictory and ambiguous explanations among the text books of physics or dynamics which deal twisting motion of a cat. Firstly, the principle of the twisting motion of a robot cat was analysed using a jointed double column model. And it was made clear analytically that a cat could restore her attitude by bending and twisting the body in the air. Then, the authors developed the robot cat which had vertebrate-type backbones and twisting mechanisms. It was confirmed experimentally that the robot cat could control its attitude by 180 degrees' rotation within 0.6 second when it was released upside-down at about 1.8m height in the air.
It has become necessary to develop a light-weight head-mounted display as a man-machine interface with a sensation of presence in tele-existence teleoperation and virtual-reality technology. Design for such displays needs knowledge of human factor and physiological optics. However, on general method has been proposed yet for the design of such displays. In this paper, we propose a method for designing head-mounted displays, and have developed four light-weight head-mounted stereoscopic full-color displays for remote operation. These displays are designed to give 3 D sensation with a sensation of reality. Two of them are very light weighted, and have wide angle of view, which are designed to teleoperate a mobile robot. The others have a link mechanism to support their weight, and have very fine resolution of view, which are designed so teleoperate a robot for fine manipulation tasks. Dynamic responses of a human operator using these displays are measured by tracking experiments. These results show which parameters of displays are related to dynamic performance. Such parameters are important factors for general design proceduer of head-mounted displays with good dynamic performance.
It is very effective for a high torque motor to make air gap as small as possible between a rotor and a stator. By polishing the surfaces of silindrical rotor and stator impregnated with epoxi resin, we developed a high torque motor of a VR type with extremely small air gaps which has no usual bearings. The rotor and stator with smooth surface play the role as a bearing. Analising the geometry of air gaps by MAGNAFIM, magnetic field analising softwere, we were be able to develop a motor with higher output torque density than that on the market.
This paper proposes a sensor using liquid conductivity among electrodes in a non-conductive spherical vessel for measuring direction-of-action changing slowly. Since the sensor is attached to a moving object, conductive liquid sealed in the vessel moves through the electrodes freely, according to posture of the vessel. The liquid dipps with multiple electrodes standing at twelve points centripetally along the inner wall of the vessel so that the conductivity among the electrodes is changed. The change is detected to calculate the normal of the liquid surface. Thus, the direction-of-action from which the liquid moves is obtained. The measurement principle and experimental results of the sensor are shown. Also, electrode's shape is analysed to show the conductivity changes in the measurement, and electrode arrangement is considered to make measurement accuracy equal inn all directions. It is confirmed that the measurement accuracy is within 7.6° in all directions.
The dynamics of a robot are highly non-linear. That it caused by the coupling torques between joints. These coupling torques result in disturbance. So the elimination of such disturbances has been one of the major problems in the robot control. On the contrary, in this paper, we propose to utilize the coupling torques for the control of the dynamics. For this purpose, we introduce a concept “dynamic redundant degrees of freedom”. It means parameters of the robot dynamic equation which can modify the dynamics task-independently. We control“dynamic redundant degrees of freedom”both to minimize joint torques in a least square sense and to keep the joint torques within their limits. The effectiveness of controlling “dynamic redundant degrees of freedom” is verified by both simulations and experiments.
At present, it is hoped that a robot which works with human or for human will be developed for various social needs. The aim of this study is in the initial development of human intimate bobot that plays board game with human. This robot is thereby called GAME-ROBOT. Its configuration consists of image processings which are based on personal computer, mecahnical arm and original robot hand. We studied the integration of these parts, their functions and the applications through the execution of Othello game. As a result., GAME-ROBOT was able to play a friendly game with human.