A virtual robot teaching system, consisting of human demonstration and motion-intention analysis in a virtual reality environment, is an advanced form of automatic programming for multi-fingered robots. We propose a new segmentation method to analyze human motion data and a virtual teaching system based on hand manipulability, in which the position and orientation of the robot hand are determined so as to maximize the robot hand's manipulability. In the segmentation method, human motion data consisting of contact points, grasp force, hand and object position, and the like are segmented into plural primitive motions and the type of task is analyzed based on the sequence of the primitive motions. A trial assembly task using a humanoid robot hand named Gifu Hand III is shown to demonstrate the effectiveness of the proposed method.
This study deals with the chip control in turning by in-situ laser heat treatment of work material. Medium carbon steel JIS S45C is chosen as workpiece material because the ductile ferrite-pearlite structure of the base material can be relatively easily changed into the hard-brittle martensitic structure by laser heat. For the purpose of practical use, the direct diode laser (DDL) is selected because of its compact size, easiness of handling and high absorbance. Prior to turning process, the continuous-wave laser beam is irradiated on the workpiece surface in a longitudinal direction. This linear heat treatment makes it possible to break chips in turning every one revolution of workpiece at which the material is embrittled. The chip breakability depends mainly on the ratio of heat-treated depth to depth of cut. In spite that the workpiece is partially hardened, abnormal wear or chipping is not observed on the tool face after turning.
Electro-chemical honing (ECH) is a hybrid electrolytic precision micro-finishing technology that, by combining physico-chemical actions of electro-chemical machining and conventional honing processes, provides the controlled functional surfaces-generation and fast material removal capabilities in a single operation. Process multi-performance optimization has become vital for utilizing full potential of manufacturing processes to meet the challenging requirements being placed on the surface quality, size, tolerances and production rate of engineering components in this globally competitive scenario. This paper presents an strategy that integrates the Taguchi matrix experimental design, analysis of variances and fuzzy inference system (FIS) to formulate a robust practical multi-performance optimization methodology for complex manufacturing processes like ECH, which involve several control variables. Two methodologies one using a genetic algorithm tuning of FIS (GA-tuned FIS) and another using an adaptive network based fuzzy inference system (ANFIS) have been evaluated for a multi-performance optimization case study of ECH. The actual experimental results confirm their potential for a wide range of machining conditions employed in ECH.
The aims of this paper are to propose a new finishing process and show the feasibility of ultra-precision finishing of thin quartz wafers used for high-frequency devices. This particular application requires simultaneous improvement of surface roughness and thickness variation with minimal material removal. Consequently, the process must allow for the control of both super-fine finishing and localized material removal. This paper proposes a method using magnetic field-assisted finishing for the finishing of thin quartz wafers. The magnetic field-assisted finishing process enables localized material removal through control of the magnetic field distribution at the finishing area; and mirror finishing is achieved using magneto-rheological fluid-based slurry. This paper describes the processing principle, the development of the finishing equipment, and the required conditions to realize the principle. The finishing experiments demonstrate the feasibility of the proposed method to perform super-fine finishing of 60 μm thick quartz wafers, in which the control of the magnetic field allows for localization of the material removal to improve or maintain the thickness variation, as desired.
This paper introduces a method of minimizing consumption in order to reduce the running cost of oxide-film CMP processes. Three types of consumables are administrated and replaced based on control limits in use time and monitored removal rate. Typically the control limits are determined separately by consumable. However the combination of consumables' use times takes various effects on removal rate. Therefore, a control limit optimization method to minimize the cost under the condition of removal rate in run-to-run lot works, wherein product lots are processed consecutively on the equipment, is developed. To achieve this, the within-wafer removal rate is modeled based on the use time of consumables, and a run-to-run simulation method to summarize consumption has been developed. The results show that this method could find a new control limit, which reduces the manufacturing cost.
In parallel mechanisms, the form and volume of workspace also change variously with the attitude of a platform. This paper presents a method to search for the workspace of parallel mechanisms with 6-DOF and 3D visualization of the workspace. Workspace is a search for the movable range of the central point of a platform when it moves with a given orientation. In order to search workspace, geometric analysis based on inverse kinematics is considered. Plots of 2D of calculations are compared with those measured by position sensors. The test results are shown to have good agreement with simulation results. The workspace variations are demonstrated in terms of 3D and 2D plots for prototype mechanisms. The workspace plots are created with OpenGL and Visual C++ by implementation of the algorithm. An application module is developed, which displays workspace of the mechanism in 3D images. The effectiveness and practicability of 3D visualization on workspace are successfully demonstrated by 6-DOF parallel mechanisms.
This paper reports on the design and simulation of a valveless micro pump which promises widely applicable in microfluidic applications. The structure of the pump comprises a PZT diaphragm and a fluidic network channel which can be easily fabricated by either conventional machining or silicon micromachining techniques. The working principle of the pump relies on the structure of the cross junction which connects the pump chamber, the outlet and two opposite inlet channels. This design allows a difference of fluidic resistance and momentum between the pump phases which results in a net flow from the inlet to outlet. The pump performance is studied by numerical simulation by applying the deformation of the PZT diaphragm. Experiments of the prototype had been conducted and working principle was confirmed.
Passive-dynamic walkers are mechanical devices that walk down a slope without being propelled by motors or controllers. In this paper we present the design of two different knee mechanisms intended for use with such walkers. At first we developed a knee mechanism that uses permanent magnets to lock the knee in the extended position and conducted walking experiments with a planar passive-dynamic walker, which was built beforehand. The walker moved down an incline and we counted the steps that it made. After performing several hundred trials we developed our second knee mechanism featuring an active release from the extended position and performed the same experiments with the same walker outfitted with the new mechanism. We compared the results achieved with the two different knee mechanisms. The active mechanism made an increased number of successful walks down the slope from which we concluded that it is more reliable and easier to use and set up than the one with permanent magnets.
To get high performance, downsizing and weight saving of the power transmission systems, the improvement of machine elements has been required. In this study, case-carburized gear materials for a high load-carrying capacity were developed. Low-alloyed steels with 1%Cr-0.2%Mo, 1%Cr-0.2%Mo-1%Si and 1%Cr-0.2%Mo-2%Ni (Cr-Mo steel, Cr-Mo-Si steel and Cr-Mo-Ni steel) were melted in a hypoxia vacuum. Test rollers were made of the developed steels, and they were carburized (Type A and Type B), hardened and tempered. Heating retention tests were carried out to investigate the softening behavior of hardness at high heating temperatures in the case of the developed steels. Roller tests were conducted under the rolling-sliding contact and high-load conditions to study the surface fatigue of the developed steels. From the obtained test results, it was found that the softening behavior of surface hardness at high temperatures in the cases of Cr-Mo-Si steel (Type A) and Cr-Mo-Ni steel (Type B) is lower than that in the cases of Cr-Mo steel (Type A) and Cr-Mo steel (Type B). In the cases of Cr-Mo-Si steel (A) and Cr-Mo-Ni steel (B), micro- and small-pitting area ratios are smaller and large-pitting life is longer than those in the cases of Cr-Mo steel(A) and Cr-Mo steel(B) under the same carburizing treatment method and high-load conditions. Furthermore, the relationship between the softening behavior of surface hardness on the heating pattern and the surface fatigue on the rolling-sliding contact of the developed alloy steels was clarified.
This paper describes the collaborative augmented reality (AR) system with which multiple users can handwrite 3D lines in the air simultaneously and manipulate the lines directly in the real world. In addition, we propose a new technique for co-creative communication utilizing the 3D drawing activity. Up to now, the various 3D user interfaces have been proposed. Although most of them aim to solve the specific problems in the virtual environments, the possibility of the 3D drawing expression has not been explored yet. Accordingly, we paid special attention to the interaction with the real objects in daily life, and considered to manipulate real objects and 3D lines without any distinctions by the same action. The developed AR system consists of a stereoscopic head-mounted display, a drawing tool, 6DOF sensors measuring three-dimensional position and Euler angles, and the 3D user interface, which enables to push, grasp and pitch 3D lines directly by use of the drawing tool. Additionally users can pick up desired color from either a landscape or a virtual line through the direct interaction with this tool. For sharing 3D lines among multiple users at the same place, the distributed-type AR system has been developed that mutually sends and receives drawn data between systems. With the developed system, users can proceed to design jointly in the real space through arranging each 3D drawing by direct manipulation. Moreover, a new application to the entertainment has become possible to play sports like catch, fencing match, or the like.
In this paper, a high rigidity and high speed rotating mechanism using a new concept hydrodynamic bearing in X-ray tube for high speed computed tomography is proposed. In order to obtain both the stability and the high load carrying capacity, the hydrodynamic bearing lubricated by liquid metal (Gallium alloy), named as the hybrid hydrodynamic bearing generates the lubricating film by wedge effect on the plane region between the spiral grooves under high loading condition. The parallelism between the bearing and the rotating body can be secured by optimizing the rigidity distribution of stationary shaft in the proposed rotating mechanism. By carrying out the fundamental design by numerical analyses, it has been made clear that the hybrid hydrodynamic bearing and the rotating mechanism are suitable for the X-ray tube used in the CT with ever-increasingly scanning speed.
Independent component analysis (ICA) is used to detect the mura regions with varying sizes and brightness levels before thresholding, then individually analyzed the mura regions in order to avoid unnecessary background effect. Defects detection is performed by partitioning test image into overlapping sub-windows and Classifying each sub-window as normal or mura region by comparing the difference of spatial distance between ICs of defective and non-defective. During the experimental process, a median filter and a high-pass filter are also respectively used to filter out the noise and enhance mura gray intensity. In this research, we developed ICA to achieve off-line learning and on-line detection.