Straightness measurement that using a laser beam is affected by wavefront aberration due to air refractive index fluctuation (air fluctuation). To compensate wavefront aberration, we propose the deep learning method. In this method, deep learning model estimate air fluctuation uncertainty and subtract it from the measurement value. In this study, we focus the change of working distance, which is the propagation distance of laser beam. The goal is to enable estimating the uncertainty by using deep learning the same as previous study.

Laser forming, a kind of non-contact bending process, has low cost and high flexibility because it doesn’t require dies or rollers. Therefore, laser forming is more suitable for manufacturing prototypes and small-batch productions than conventional processes. However, the complexity of deformation mechanism makes it difficult to optimize the process conditions for the final shape. In this study, the feed-forward neural network (FNN) is trained by teaching datasets to build a model predicting the scanning paths required for the final shape of a metal sheet. To collect the datasets for machine learning, a large number of imaginary datasets are prepared instead of the experimental ones. In addition, for verification of the trained FNN with imaginary datasets, the experimental datasets are employed to evaluate the prediction performance by the model. Obtained results show that the imaginary datasets can train the model to have a prediction accuracy as high as the experimental data.

Laser welding is widely applicated in the precision manufacturing field. Metal irradiated by femtosecond or nanosecond laser will induce many nanostructures around the laser-irradiated area, which will affect the connection of two metals or the surface properties. Therefore, it is necessary to further study the mechanism of the laser-induced nanostructure’s generation. In this study, we try to employ high-speed imaging to observe the laser irradiation process on the nanoseconds scale. The image intensifier and high-speed camera were used, and the signal was controlled by the delay generator to capture the plasma plume at different times after laser irradiation, and the plasma plume formation and disappearance process was observed.

We develop an alignment monitor which can work in a severe environment of high magnetic field and vacuum highly stably for term longer than several months. Here, we consider an optical simulation model for the alignment monitor consists of ball lens. The model is based on a combination between an approximation model for collimated beams by using single mode optical fibers and a Gaussian optics model. We evaluate validity of the model by comparing estimated results by using the model and experimental results.

In this study, a technique is proposed to realize the detection of the micro-contacts. The technique for detecting the contacts is based on the principle using transfers between sputtered thin films coated on both of contact surfaces. The micro-contacts between both surfaces of a rough ceramics ball and a smooth glass flat are observed and measured, using the proposed technique. After indentation, surface images of contacts were measured on the glass surface by an optical microscope and SPM. The result showed that it is found that micro-contacts of less than submicron in lateral size can be clearly detected by using the technique.

The National Metrology Institute of Japan (NMIJ/AIST) has been developing a two-point diameter calibration technique of spheres using micro-coordinate measuring machine (µ-CMM) with a low-force touch probe. In this paper, an approach for reducing the interaction force between the sample surface and the probe of the µ-CMM will be reported to reduce the measurement uncertainty of sphere diameter. Compared to the previously used probe with a tip radius of 500 µm, T-shaped styli with a tip radius of about 8 µm was developed. This paper describes the relationship between interaction force and µ-CMM probing performance with stylus tip radius sharpening.

Quantum sensors or computers using cold atom are attracted attention in recent years. An ultra-high vacuum environment of less than 10-6 Pa is required to handle cold atoms. To contribute miniaturization of quantum devices, we have been developing a miniature ion pump with a high-efficiency plasma source that can evacuate the cold atom generation chamber down to the ultra-high vacuum environment. In this report, we present the vacuum evacuation characteristics obtained by measuring the discharge current of the ion pump.

Germanium attracts attention as high frequency device material because it has high electron and hole mobility. However, as it has low thermal conductivity it suffers heat dissipation problems. Therefore, we considered that combining a germanium substrate with a diamond heat spreader substrate enables efficient heat dissipation. Conventional hydrophilic bonding using oxidizing treatments has been used for compositing substrates. This process may form an intermediate oxide layer which can be a thermal barrier. Thus, we used hydrophilic bonding using reducing treatments in this study. In this process the germanium substrate is immersed in hydrochloric acid. It is possible to functionalize the germanium substrate and remove the native oxide layer at once. It is supposed that this process can make the intermediate layer thinner and improve heat conduction.

Electrocardiogram (ECG) detection is beneficial for medical monitoring of patients. In order to achieve a stable ECG signal detection, the electrode is made of copper wire textile, and then the circuit suitable for the electrode is designed according to the principle of capacitive coupling. To receive a more stable signal, a color sensor is used to estimate motion artifact caused by small movements of the human body. In breathing state, the ECG signal with obvious QRS-T waves were measured by the woven electrodes, and the artifact signal was measured by color sensor. After processing with adaptive filtering NLMS algorithm of the two signals, the deviation of the ECG signal is suppressed, and the SNR increased by 5.3dB.