In recent years, the health of many people is damaged by asbestos. Therefore, to detect very small quantity of asbestos that exists in general environment is very important. According to an officially fixed way, existence of asbestos in a certain building material is discriminated according to whether there are more than 4 crystals of asbestos per 3000 particles in samples made with crushing a part of the building material. The particle counting is performed manually, therefore it needs great time and labor. For automation, some studies to count particle with image processing are performed. To ensure consistency with conventional manual method, the automated method should be able to obtain equivalent outcome to manual methods. In this paper, we propose an automatic particle count method based on multiresolution analysis. At first, the classifier which can distinguish existence or nonexistence of a particle at the center of ROI (region of interest) is constructed. Next, the image is scanned with the classifier on multiresolution, and small areas containing a particle are found. To construct the classifier according to the counting result by manual method, we can achieve the result as same as manual one by the proposed method. To confirm validity of the proposed method, we did some experiments to apply the method to the images that particles are identified manually. As the result, we confirmed that 10% of false positive and false negative are contained respectively. Because the proportions are roughly same, the total count is as same as manual counting.

View full abstract

In the field of semiconductor exposure machines, air-type active anti-vibration apparatuses are widely used. This apparatus equips an actively controlled air spring which has a pair of anti-resonance and resonance on frequency response. Due to this characteristic, it is known that deformation of main body supported the air springs occurs because of the difference of pressure in each air spring. In order to restrain the deformation of main body and to control their generated forces, the pressure feedback was especially implemented as a minor loop for the main position and acceleration loops. However, it is gradually undesirable to install pressure sensors in the field of industrial application because it takes cost and the sensors have low resolution for the management of the generated force. Then, this paper proposes an identity observer that estimates inner pressure of the air spring and shows that the estimated pressure signal can be used as a feedback one. Firstly, a tunable block diagram, which is equivalently transformed from the formal identity observer, is shown. Next, tuning results based on both frequency and time responses are shown and then feedback results using the estimated acceleration and pressure signals are successfully shown.

View full abstract

This paper introduces an object detection method using features described information about a concavo-convex shape of an object that are obtained by using a small camera that controls the illumination direction. An image of the binarized gradient feature containing information about the shape of the object is generated by the Coded Multi Flash Imaging. The proposed method can reliably detect a texture-less object by using this feature image in the matching process. In addition, processing time for matching is significantly reduced by using a small number of pixels. Experiments using 200 actual images confirmed that the proposed method achieves a 98.5% recognition success rate and a 607 msec processing time.

View full abstract

In optical lithography, the projection optics is known as the high accurate exposure tool for 3-dimensional device patterning such as MEMS (Micro Electro Mechanical Systems). Required optical performances of the projection optics are resolution capability of pattern density and depth of focus (DOF) on step-heights. Generally, larger numerical aperture (NA) of projection optics is quite effective on higher resolution, but loses DOF on the contrary. The capability of micro-lens-array (MLA) patterning has the resolution limit around 20-30μm in Sag (lens height) by the conventional high NA projection lens. Therefore, projection optics with lower NA and deeper DOF is required for the MLA patterning with 100μm Sags for various applications. This paper reports the effectivity of the combination of the newly developed projection lens with adjustable NA of 0.06-0.22 and a photo-mask with quasi-continuous intensity distribution by a density control of scattered square micro-dot-patterns. By experiments, this method has been proved to realize the MLA patterning with 100μm Sags with errors less than surface roughness of 2μm RMS and Sag shape deviation of 5μm.

View full abstract

In semiconductor industry, production of ever flatter, thinner and larger Si wafer are required to fulfill the demands in high integration and cost reduction. A severe problem encountered in wafer thinning process is the warp and distortion of wafer induced by the residual stress and subsurface damage. Chemo-mechanical grinding (CMG) process is emerging process which combines the advantages of fixed abrasive machining and chemical mechanical polishing (CMP), offers a potential alternative for stress relief. However, since the sodium carbonate is used as additives in previously developed CMG wheels, the metal ions like sodium ion were possibly introduced on the machined wafer surface, and could risk in metal contaminations. In order to develop the high quality CMG wheel excluding sodium carbonate, CMG wheels with different recipe were developed. So far, CMG has been studied to investigate the effects of wheel geometry and grinding conditions on the removal rate, and effects of bond materials and additives on the surface quality in our previous researches. In this study, the focus was given to the wheel mechanical property (bending strength, elastic modulus) of CMG wheels. The difference in wheel properties is discussed by association with CMG performance including wheel wear and surface roughness. The result have demonstrated that the surface roughness of newly developed CMG silicon wafer is equivalent to that of CMG containing sodium carbonate as additive silicon wafer without any grinding marks.

View full abstract

This paper discusses optimal control of jump motion for die-sinking electrical discharge machining(EDM). Efficient flushing machining debris from a clearance between an electrode and a workpiece is important to improve machining stability, precision and speed. The fast jump motion realizes the efficient flushing and decreases the idling time, but excites the machine vibration. Moreover, reaction force from dielectric fluid causes the machining unstable. In this paper, optimal control has been developed for the jump motion. The first controller was designed using 2-inertia model with optimal control theory whose objective function included vibration and speed. The second one, which consisted of sine and square waves, was derived from Fourier expansion. Added to these, the disturbance reduction control strategy has been proposed for the large or thin electrode. The new motion control improves the jump speed and reduces machine vibration and electrode damage, and it has been confirmed the efficiency in experiments.

View full abstract

In case of the five-axis controlled machining, there is interference problem among tool, workpiece, jig, machine structures such as table and spindle. In this study, these 3D models were loaded to CAM, and, moved to position corresponding to candidates of all tool postures on tool path. Then, by performing interference judgment, candidates of tool postures without interference were calculated. Therefore, tool path without interference was generated. When interference judgment is performed, process time was shortened by parallel processing using GPU and polygon model. Additionally, this CAM was corresponded to all types of five-axis controlled machining, table-tilting type, table/spindle-tilting type, and spindle-tilting type, and, improvement of general versatility was conspired. Finally, as the result of the machining simulation using all types machining, the effectiveness of the devised technique was confirmed.

View full abstract

This study deals with a novel generation method of tool paths in a main-processor of a CAM. Currently, some tool path generation methods such as contouring path, scanning path, spiral path and constant scallop height path are installed in several CAM applications. However, since these tool paths are generated by differential operations and convergence calculations, complex branches on conditions are required, and the operation is unstable behavior at extremum neighborhood. It still threatens to cause clash of the operation. Therefore, two parameters (u, v) composing a machined surface are the focus of this study. In our proposed method, machining points are calculated on a plane constructed by the parameters. Then, since the calculation is executed in response to the real space, a homogeneous tool path is obtained. By using our proposed method, it is possible to generate tool paths quickly and effectively without the extremum problem and complex branches on conditions. This paper explains the particulars of our proposed method and the implementation method for two spiral paths. Finally, the effectiveness of our method is confirmed with machined workpieces.

View full abstract