Epitaxial thin films of a ferroelectric perovskite-type oxide grown on single-crystalline SrTiO₃ (100) were transferred onto a flexible printed circuit (FPC). In the case that the thin films were directly adhered onto FPC using a copper foil double-coated conductive adhesive tape (Cu double-sided tape), serious cracking and exfoliation occurred during the transfer process. To avoid these damages, we have tried to insert a metal buffer layer with excellent ductility between the ferroelectric oxide thin film and the Cu double-sided tape. The platinum buffer layer was found to be appropriate to establish a crack- and exfoliation-free transfer process.

The interaction between diamond surfaces and alkali metals were studied experimentally and computationally. Li reacted with diamond above 600°C and significant etching of (100) surface was observed. An attempt was made to apply this etching reaction for electrical discharge machining of diamond tools.

In order to establish a method to precisely predict penetration depths for macro and micro laser welding at manufacturing sites, we have investigated how to determine experimental conditions for Hablanian plot. We need to obtain three approximate curves (standard approximate curve, error approximate curve, and coefficient approximate curve) for an accurate prediction of penetration depths. We have discussed how to obtain three approximate curves using laser welding data by the use of the fiber laser. As a result, the following findings could be derived. (1) Cubic polynomial is probable as one of the candidates of a function for three approximate curves. (2) A focusing lens used and a spot diameter should be first determined. Second, laser welding experiments should be carried out with changing Log₁₀(Vd/K) and P at regular intervals. (3) Prediction errors can be respectively reduced down to ∼0.06 and ∼0.03 for macro and micro welding above a penetration depth of 4 mm and 0.4 mm.

A copper has been widely applied for various industrial components such as heat exchangers, heat pipes, electro circuits and motors because of having excellent properties such as thermal conductivity and electrical conductivity. Besides the copper is expected to be used for handrails, doorknobs etc., which touched by many people to prevent from infectious diseases because of having the excellent properties such as an antibacterial property and a virus inactivation property. However, copper is low strength to only about 40% as strong as stainless steel. To add a new function of virus inactivation property to the surface keeping the strength, it is necessary to develop a coating technology to form a copper layer on the surface. Thus, we have developed a multi beam type laser metal deposition method with a high intensity blue diode laser which newly developed with an output power of 200W. As a result, Cu-Zn alloy layer was formed on the stainless-steel plate with no pores at the output power of 30W and scanning speed of 6mm/s.

In order to compensate the non-linear characteristics in many kinds of electronic devices, the exponentiation conversion circuit that can change the power exponent to any value has been proposed. The exponentiation conversion circuit multiplies the logarithmically converted input signal by the power exponent value to perform exponential conversion. As a result, we can obtain the power function characteristic of the power exponent value. In this paper, we proposed a transconductance circuit that can increase the voltage gain, which decreases with lower power supply voltage. By parallelizing the input and output signal voltage windows with the transconductance circuit, the maximum exponent value was secured to be 2.9, which is equivalent to the conventional value, furthermore the signal dynamic range was 42.7 dB in simulation results. By evaluating its prototype IC with 0.6 µm CMOS process, we confirmed the signal dynamic range was 30.0 dB and maximum power exponent value was 3.4.

The novel wireless system that distributes polarization of the electromagnetic wave using conventional device which enables PSK modulation is proposed. The system achieves high quality wireless communication by using suitable polarizations both of transmission and reception waves. This technology leads longer communication range than that of conventional 5G wireless systems under a non-line-of-sight environment, and therefore, it enables to reduces the introduction cost of the wireless network useful for Internet of Things applications.

Non-invasive positive pressure ventilation (NPPV), an artificial respiration therapy, is a treatment in which a mask is attached to the patient’s face and delivers gas into the mask to support breathing. The NPPV mask straps are required to attach and secure the mask in the appropriate facial position, but the tension strength of the strap is adjusted by the sensation of the hands. The strap uniformity and fine-tuning strap tension are judged by the skill of the operator and the amount felt by the finger. In the future, additional strap operation and adjustment methods will be required to meet the needs for reducing the burden on the face.

In this study, we fabricated a mechanism that can control the fine adjustment and fixing of the tension of the mask straps. A small amount of strap tension can be adjusted by operating the shaft. This makes it possible to control slight tension that is difficult to grasp with the sense of the operator’s hand. In addition, this mechanism allows the operator to control the strap while controlling the movement of the mask body. This leads to the establishment of a mask fitting method suitable for each patient. The developed stabilizing technique simplifies operator mask fitting, allows strap operation and adjustment, and reduces facial strain.

This research aimed to compare of accuracy for machine learning using pulse wave. The subjects were 32 healthy young adults. They were divided to two groups by psychological tests. The pulse waves were measured during four emotional audiovisual stimuli. The subjects were discriminated into the mental stable or the mental unstable by Multi-Layer Perceptron (MLP) and Recurrent Neural Network (RNN) by using pulse wave, and the accuracy was calculated. The rate of the RNN was higher than that of the MLP for the most of the stimuli. These results suggest that RNNs would suitable for machine learning using pulse wave.

Efficient equipment operations by using energy demand forecast and optimized operation schedule have become common. To reduce the influence of forecasted errors, it is effective to apply re-forecast methods that updates the forecasted value using the actual demand at the last minute and re-schedule methods. Therefore, this paper shows these effectiveness and limitations based on the quantitatively evaluation for office buildings with heat storage facilities. We first propose a novel forecast method with re-forecast. Next, we have performed numerical simulations of equipment operation including re-schedule to clarify the quantitative effects of re-forecast and re-schedule. Finally, through the analysis of these results, we have identified issues for further improvement of energy utilization efficiency.

In recent years, projection mapping has become popular in the commercial and artistic fields. Among them, there is projection mapping that the viewer's motion and the movie interact. They are called interactive art, and are attracting attention. On augmented reality, pattern hiding methods have studied to estimate depth using only a camera and projector. However, there are color breaking at the border of the grid and flicker occurs when pattern change. Therefore, in this study, we used the phase shift method for interactive art, and investigated the possibility of depth estimation. As a result, we calculated phase image and reduced color breaking and flicker. The errors in the phase images, it was found that the accuracy was sufficient to recognize an adult.

Tackling diabetes, an increasingly common lifestyle-related illness, requires information on patients’ lifestyle and habits. Blogs maintained by patients afflicted with the incurable illness may be useful for analyzing how lifestyle affects health-status. In this paper, we propose an intrinsic expression extractor that extracts keywords related to lifestyle and health-status from blogs of patients diagnosed with type-2 diabetes. To counter class imbalance and add to the corpus, the proposed method tags each keyword with information based on cue-words extracted from manually tagged data. The named-entity recognition (NER) for the extracted keywords uses a bidirectional gated recurrent unit neural network (BiGRU) and was evaluated for accuracy by cross-validation. We obtained F1-score of approximately 0.76. Although the accuracy of extraction can further be improved, the novel approach has applications in analyzing and improving the lifestyle of diabetes-afflicted patients.

Adam is one of the general optimization algorithms in neural networks. It can accelerate convergence speed while learning. It has, however, two problems. The first is that the final performance of a network trained by Adam, such as generalization ability, becomes worse than the one trained by SGD, in applications to large-scale networks. The second is that values of the learning rate tend to be large at the early learning stage; as a result, the values of network parameters, such as weights and bias, become too large by a first few iterations. In recent years, research has been conducted to solve these problems. AdaBound has been proposed for solving the first problem. This is a method switching dynamically from Adam to SGD. RAdam has also been proposed, for solving the second problem. This applies a method called WarmUp, which sets a small learning rate at the early learning stage and gradually increases it, to Adam. In this study, we propose to apply WarmUp to the upper limit of AdaBound's learning rate. The proposed algorithm prevents parameter updates at extremely large learning rates in the early learning stages. Therefore, more efficient learning can be expected than the conventional method. The proposed method has been applied to learning of some types of networks like CNN, ResNet, DenseNet and BERT. The results show that our method has improved performance compared to the traditional method, and an image classification task has shown a tendency to be more effective in large networks.

Air quality monitoring (AQM) at fine and granular levels helps in implementing better air pollution control and mitigation plans. Low-cost sensing is a relatively new paradigm for AQM at high spatial and temporal resolutions. However, the data obtained in this technique is less reliable due to various error sources involved. Calibration is a promising approach to counteract error sources and to improve the accuracy of the data obtained from low-cost sensors (LCS). However, fitting a calibration model that works for every application is cumbersome since each application is associated with varying sources of error. Hence, we focus on finding a better calibration model for a given application. Our approach is application-centric. We classify LCS applications into static and mobile, based on sensor deployment in the field. Then, we sub-categorize both stationary and mobile applications further to narrow down to understand each case. Next, we analyze the possible error sources in each application and the corresponding techniques to counteract them. Then, we map applications to the calibration models via possible error sources and assign weights/scores to each parameter based on our analysis. Our quantitative analysis determines scores for each calibration model which results in finding the best suited model for a given set of conditions and applications based on the highest score. We verify our analysis with real time data obtained from LCS deployed in Chennai city, India.

The objective of the present study was to evaluate the effect of extremely low frequency (ELF)-electric field (EF) treatment on decubitus ulcer. Hairless mice (Hr-/kud) were used, and ulcers were formed by magnetic compression of the back skin. The extent of the ulcer was assessed by the diameter of the wound and the crust area. The observation period was 14 days from the day of pressure. Three experimental groups were assigned: a group treated with EF from before the formation of the pressure ulcer until the end of observation, a group that started EF treatment on the day of magnet compression, and a control group that only had pressure ulcer formation. EF was generated at 50Hz at 10kV/m between parallel plate electrodes. EF treatment on the mice was 60 minutes per day. Results showed that the maximum wound diameter was reached in all groups for 3 days. Both decubitus ulcer diameter and crust diameter in the field-treated group were smaller than those in the control group, suggesting an inhibitory effect of ELF-EF on the process of decubitus ulcer development and repair.