We propose a mechanical hand for grasping washed clothes. The final aim of this research is to perform the task of laundries arrengement. The task needs some patterns of hand's operation and manipulator's movement. In this report we examine to automate separation of clothes from mounted clothes including several kinds of laundry and grasp of edges of those clothes.
A control algorithm for transporting a large object by cooperation of multiple velocity-controlled mobile robots is discussed. Passive joints are installed into each robot as mechanical compliance to avoid excessive inner forces caused by mutual positioning errors among robots. Two cooperative systems consisting of three mobile robots are designed as cooperative mechanisms and their characteristics are analyzed from the control point of view. After choosing one type with more suitable performance for feedback control, a prototype system is developed. Experimental results of cooperative transportation with the system show the effectiveness of the proposed system.
Parallel wire driven systems have many advantages such as high speed and safety to humans. When the motion of the system is restricted less than 6 D.O.F., mechanical constraints are utilized to fix the unnecessary D.O.F. and the remained freedoms of the system are driven by the parallel wire mechanism. The mechanical constraints should be light for realization of high speed or safety to humans. As the results, the stiffness of the directions made by mechanical constraints decreases and oscillation occurs in many cases. In this paper, we propose a stiffness increase method basing on nonlinear stiffness of the mechanical constraints and the wire drive units. In the proposed method, interference of wire tension with mechanical constraints must be effectively used. In order to obtain an appropriate interference pattern, an algorithm to design wire directions is also proposed. Finally, the effectiveness of the proposed method is demonstrated through several experimental results.
Electric powered wheelchairs (EPWs) are supposed to be widely spread in the near future. Many models are now available in the market, but unfortunately, it seems most of them are 'difficult-to-drive' because of their manipulating sensitivity. In our previous study, it was shown that the improvement of the manipulability of an EPW could be accomplished by adjusting its 'manipulating impedance' to that of the users' preference. In this article, an alternate and a considerably simple solution of the problem as stated above is introduced. When designing a control system for an EPW, one has to consider (1) realization of the manipulating comfort, and (2) treatments against parameter variations at the same time. The idea employed for assuring manipulating comfortability is the same as our previous study, but the method adopted for assuring robust performance against parametric variations is completely new. Key theorem and its proof will be given, that should form a theoretical basis of this article. Finally, we must say that the design method proposed here should not be restricted to the design of the control system for an EPW.
This paper describes a control system for the heavy object handling manipulator. It has been developed for the blanket module remote maintenance system of ITER (International Thermonuclear Fusion Experimental Reactor) . A rail-mounted vehicle type manipulator is proposed for the precise handling of a blanket module which is about 4 tons in weight. Basically, this manipulator is controlled by teaching-playback technique. When grasping or releasing the module, the manipulator sags and the position of the end-effector changes about 50 [mm] . Applying only the usual teachingplayback control makes the smooth operation of setting/removing modules to/from the vacuum vessel wall difficult due to this position change. To solve this proper problem of heavy object handling manipulator, we have developed a system which uses motion patterns generated from two kinds of teaching points. These motion patterns for setting/removing heavy objects are generated by combining teaching points for positioning the manipulator with and without grasping the object. When these motion patterns are applied, the manipulator can transfer the object's weight smoothly at the setting/removing point. This developed system has been applied to the real-scale mock-up of the vehicle manipulator and through the actual module setting/removing experiments, we have verified its effectiveness and realized smooth maintenance operation.
A leg-wheel robot has mechanically separated legs and wheels, and it performs high mobitity and stability on the rough terrain. In the first part of this paper, we propose a gait algorithm which enables the robot to preserve continuous locomotion under the random velocity command such as the human operator inputs. The gait algorithm, which uses predictive control, determines the timing of the legs' lifting to avoid the legs reaching the border of the work space. The predictive control is based on the comparison between the time necessary to return the support legs and the time left before they reach the border of the work space. In the second part, a velocity limitation method for the gait algorithm is discussed. This method limits the velocity command when it exceeds the mechanical perfor-mance of the robot. Combined use of the velocity limitation method with the gait algorithm ensures the continuity of locomotion, and makes the efficient gait pattern with a long step length and low frequency of leg phase change. The proposed algorithms are evaluated by simulation.
A new vehicle assembly for holonomic and omnidirectional mobile robots is presented. A caster-drive mechanism is one of the feasible solutions to allow a holonomic omnidirectional vehicle to equip standard wheels for its driving wheels. A traditional synchro-drive transmission is applied to the caster-drive system. Multiple drive-caster wheels are mechanically coupled and simultaneously driven and steered by respective motors. A rotational stage, orienting a vehicle frame, is attached on the center top of the mobile base. Position and translational velocities of the mobile base are controlled by the synchronized drive-casters. The orientation of the vehicle frame is controlled by the rotational stage completely decoupled from the translational motion of the mobile base. The synchro-drive transmission brings enormous advantages to the holonomic vehicle; three motor driven, all wheel drive, decoupling control, and simpler mechanism. In addition, an offset gear drive is also applied to the drive-caster wheel. This gear transmission enables a decoupling control of the drive shaft from the steering shaft on each wheel. By applying those mechanisms, the vehicle kinematics becomes very simple. Kinematics and statics of the proposed omnidirectional vehicle are analyzed and verification of the mechanism is shown by computer simulations.
This paper proposes a method for constructing a decision tree and prediction ones of the landmarks that enable a robot with a limited visual angle to make decisions without self-localization in the environment. Since global positioning from the 3-D reconstruction of landmarks is generally time-consuming and prone to errors, the robot makes decisions depending on the appearance of landmarks. By using the decision and the prediction trees based on information criterion, the robot can achieve the task efficiently.