This paper presents new concepts to apply Self-Organizing Maps (SOM) to vehicle path problems. It is reported SOM is capable of solving a traveling salesman problem (TSP), one of the vehicle path problems. However, it is not investigated how SOM is applied to other types of the vehicle path problems, such as the shortest path problem (SPP) and n traveling salesmen problem (n-TSP). Numerical experiments prove that SOM cannot lead to good solution when it is applied to SPP and n-TSP. To improve SOM, two multi-neuron concepts are introduced to solve these problems. Numerical experiments verify that SOM using multi-neuron leads to better solutions than the conventional SOM does. Furthermore, SOM with variable number neurons is proposed to make SOM's solution converge efficiently.

抄録全体を表示

The ultrasonic motor (USM) has excellent performance and many useful features that electromagnetic type motors do not have. It has been used in many practical applications. A characteristic of the USM that is affected by friction is strong nonlinearity, which makes it difficult to control. This paper proposes a position control method for the USM using Support Vector Regression (SVR), which is a regression method for Support Vector Machines. It is a newly proposed method of machine learning that does not have the disadvantages of Neural Network such as a large number of learning times, local-minima, overfitting and so on. The proposed method uses an SVR controller combined with a PI controller. The SVR controller performs nonlinear input-output mapping of the USM. The learning of the SVR controller uses training data obtained from experiments. The effectiveness of the proposed control method is confirmed by experiments.

抄録全体を表示

A fast two-dimensional measurement system of birefringence dispersion is developed for inspection of LCD retardation film. An optical arrangement of this system consists of a polarizer, a carrier retarder and an analyzer. The carrier retarder can be introduced a carrier frequency into an intensity distribution along wave number. A Xenon flash lamp is employed as a white light source and synchronized with the line type of imaging spectrometer. The detected intensity distribution changes the sinusoidal wave along wave number. A phase distribution is calculated by Fourier transform method. The retardation distribution of a polymer sample is analyzed to subtract phase distributions with and without the sample. In the experiments, the some polymer films are measured as demonstrations and evaluated to the non-uniformity of retardation distribution.

抄録全体を表示

A low-pass filter is almost always applied to data of finite length. Therefore, unexpected vibration occurs at the data end of the low-pass filter output. To solve this problem, the previous report of this paper has proposed that two types of low-pass filters (i.e. spacial and frequency) can be used in a point symmetric extension method. To make the above proposal more specific, a point symmetric extension method was applied to reduce an abnormal value occurring in the filtered data at the data end where the center of the point symmetric method is applied. Before that, the data end was replaced by the generalized low-pass filtered output. Consequently this paper proposes a method to determine the center point of this point symmetric extension based on the sum of squared errors between the primary and the filtered output. This new method also uses a spatial low-pass filter and a frequency low-pass filter. The method is applicable not only to a low-pass filter that uses a point symmetric extension method but also to all other types of low-pass filters for minimizing the sum of squared errors between primary and filtered data.

抄録全体を表示

This paper describes a new approach for the detecting pattern edge positions and roughness of LSI ultra-fine patterns. As LSI pattern lines become narrower, exact estimation of line edge roughness and line width roughness as well as edge positions, becomes indispensable. The proposed approach comprises the function approximation of cross-sectional signals at each edge in a SEM image and the automatic determination of the optimum edge detection region used in the function approximation. The approach enables us to detect edge positions and roughness with high accuracy and high reproducibility.

抄録全体を表示

This paper describes the experimental results on the error inducing factors in the high-precise surface profile measurement by a laser confocal range sensor. One of their factors is tilt angle of a surface at a measuring point. According to the experimental results, the more the tilt angle increases, the more the measurement error increases, regardless of the nominal measuring distance. The possible cause of this factor is the existence of multiple peaks in the reflection spectrum of laser beam. Although the prime peak reflection indicates the just confocal position, it may cross over with the second peak reflection or include some noise reflection in a case. For suppressing the error inducing factor described above, the filtering for the reflection spectrum is presented in this paper. The experimental result clarified the relationship between measurement error and the threshold value of filter and confirmed that the standard deviation of measurement error by the filtering decreases over 30%. Finally the conclusion including the design remarks on the spectrum filtering and further study are mentioned briefly.

抄録全体を表示

Ultrasonic vibration has been applied on micro drilling of hard and brittle materials such as single crystalline silicon and glass etc. and has leaded to superior results. These methods proposed so far, however, need the special equipments to vibrate tool or workpiece. This paper proposes a new ultrasonic vibration method that vibrated the cutting fluid in which tool and workpiece are soaked. This principal is to accelerate the flow of chips by the cavitation generated in the fluid. Micro drilling of SUS304 and aluminum alloy is taken up and the drilling performance is discussed. In addition, the behavior of cavitation is observed with a high-speed camera in drilling of acrylic resin. The obtained results are as follows. Compared with conventional drilling, this method reduces adhesion and/or welding of chip owing to the cavitation generated at the front of the drill and in consequence brings out longer tool life, more stable chip removal and smaller surface roughness. Furthermore, the cooling effect is also superior in various coolant methods.

抄録全体を表示

Cutting experiments were conducted on an NC precision cutting apparatus to determine relationship between amount of oil particle on the precut surface and lubricant applying effect. Minute diameter-oil particles could be obtained by accumulation of oil mist which was generated by a spray gun. The amount of the oil particle could be controlled by the accumulation time. Reduction percentage in the cutting forces surely increased with increase of the accumulation time. AFM and ellipsometer analyses indicated that the minimum thickness of the oil film on the precut surfaces which could produce the maximum reduction percentage was approximately 2nm.

抄録全体を表示

The purpose of this research is to develop a micro electrical discharge machining using electrostatic induction feeding. This method can reduce the minimum discharge energy per pulse eliminating the influence of the stray capacitance in the circuit. Since non contact electric feeding is possible, this method enables EDM using a high speed rotating spindle or air bearing spindle. Gap voltage calculated from the equivalent circuit of this method agreed with the experimental results. It was thus found that the electric discharge energy per pulse can be decreased by shortening the feeding electrode length. Surface roughness of Rz 0.24μm and micro wall thickness of 1.0μm were obtained using a normal electrical discharge machine, which is difficult even with EDM machines specially designed for micro machining.

抄録全体を表示

Highly flatness is necessary for silicon wafers. An optimization method of polishing conditions of Double Sided Polishing (DSP) was proposed for the purpose. A wafer is held in a carrier of a DSP apparatus. The carrier moves in conjunction with the movement of a sun gear and an internal gear of the DSP apparatus. Polishing pads that contact with both sides of the wafer rotate independently with the wafer and gear movements. The wafer motion in the DSP apparatus was modeled in order to analyze the rotation speed and the relative velocity to the pads. The DSP conditions such as rotation speeds of the gears and the pads in the DSP apparatus were optimized in order to unify the friction distance distribution in the wafer. The Quasi-Newton method was used for the optimization. A penalty function was also used to take apparatus limitations into consideration as constraints in the optimization.

抄録全体を表示

In optical electronics systems, particular optical devices such as mirrors, lenses, diffraction grating, etc. play respective roles. Therefore, many optical devices are combined to consist of a optical system. This causes the problem that the arrangement error of these devices influences the total accuracy of optical system. In addition, new optical devices with multi-function, e.g. diffraction and focusing, have been recently required to miniaturize optical systems and to decrease cost. However, their shape is generally very complicated. Typical example is such a shape as curved microgrooves on a 3-dimensional complex surface. To machine the shape, milling or turning by use of a rotational cutting tool is not suitable for tool interference. In the study, a non-rotational cutting tool is thus employed. In case of machining microgrooves on sculptured surface with a non-rotational cutting tool, 6-axis control is necessary to control tool attitude and tool direction. The study deals with the development of CAM system for 6-axis control ultraprecision grooving. From the experiments of creating microgrooves on a spherical surface and a sculptured surface, it is found that microgrooving on their surfaces is possible by 6-axis control.

抄録全体を表示

The 5-axis control machining has a lot of advantages in the machining of complex shapes. The major problem in 5-axis control machining is the tool interference with workpiece. Some commercial 3-dimensional CAM software are already offered, which consider the tool interference problem and allow us the interference-free 5-axis control machining. However, the policy of above CAM software to the tool interference avoidance is that the tool attitude is modified only at where the tool interference is detected. Thus, the tool attitude suddenly changes with the redundant movement. Because such movement cause the over and under cutting, it is necessary to be avoided as much as possible. We have already proposed the tool path generation method for 5-axis control machining using an original configuration space. In the previous paper, the interference-free cutter location data with the least amount of the change in tool attitude was generated while referring to all of the configuration space information at several cutting points. In this study, tool attitudes after interference avoidance are decided by referring to configuration space information for the whole of machining surface. The technique of the previous paper is enhanced, the continuousness of the tool attitude change in the direction of not only tool feed but also pickfeed is considered, and the method for generating the cutter location data for the optimum 5-axis control machining is developed. As a result, the usefulness of the devised CAM system is experimentally found.

抄録全体を表示

This paper proposes a new micro electrical discharge machining method which can machine micro rods and holes quick and cheap without a special equipment. With this method, a rod electrode with negative polarity is rotated and fed into a plate electrode to form a hole. After the rod electrode returns to the initial position, the axis of the rod electrode is off-centered from the center of the hole at a certain distance. The polarity of the rod electrode is then reversed, and the rod electrode is fed into the plate electrode, either with or without rotation. Since holes are formed by the rod electrode itself, the rod electrode can be machined very precisely. Since this method does not need initial positioning of the tool electrode with respect to the plate electrode, the operation is easy and short. However, a straight and long rod electrode is difficult to be formed because of the plate electrode wear. Moreover, when the axis of the rod electrode does not coincide with the axis of rotation, the rod electrode cannot be machined precisely. Thus the two step method, run-out measuring method, and dummy hole diameter measuring method were newly proposed to obtain more improvement of accuracy. High accuracy axis was formed using these methods even when the rod electrode does not coincide with the axis of rotation.

抄録全体を表示

This paper describes a micro-magnetic bearing (MMB) motor, in which, for the miniaturization, the rotor is supported by a one-axis-controlled magnetic bearing and the axial displacement is measured using capacitance variation between electromagnet cores and the rotor to stabilize the system. To drive the MMB rotor, an induction motor is chosen from the viewpoint of easy-to-integration. The induction motor is designed so that influences of magnetic and electrical interferences on the magnetic stiffness and the accuracy of the measured axial displacement can be reduced. The experimental results for an MMB motor with 6-mm-diameter rotor show that the rotor could levitate and rotate at 10,000 rpm stably. Magnetic stiffness of the MMB in the radial and angular directions were 1.32 N/mm and 0.64 Nm/rad, respectively and vibration amplitude of the rotor in the radial and angular directions were 2.3 μm and 3.5×10^-4 rad at 5,000 rpm, respectively.

抄録全体を表示

Microspike-arrays are formed on a tungsten surface by irradiating femtosecond laser through a mask aperture in the air. Formation mechanism of such microspike-arrays was discussed with analysis of laser light diffraction. Intensity distribution diffracted through a mask aperture was calculated by using the Fresnel-Kirchhoff integral formula, not the Fresnel diffraction formula, because the distance between a mask aperture and a tungsten surface is almost as short as the size of a mask aperture. Assuming that the ablation rate depends on the diffracted intensity, ablation process by repetitive irradiation was analyzed. Estimated ablation results were consistent with the experimental results. Microspike-arrays have potential in application for such as emission cathodes in a Field Emission Display (FED).

抄録全体を表示