Cancer has been the leading cause of death among Japanese since 1981, and many people die from it every year worldwide. While various measures have been taken to reduce the mortality rate of cancer, circulating tumor cells (CTCs) in the blood have been attracting attention in recent years. In the past, CTCs were detected by visual inspection by a physician or by an expensive machine, but these methods required much effort by the physician and required only EpCAM-expressing cells to be detected. In addition, detection by image processing has been used, but it has the problem that the area of interest is only a part of the area and there are many false positives. In this paper, we propose a two-step classification method that focuses on the shape and surface of cells. In the proposed method, multiple shape and surface features are obtained for four types of cells in blood images: Clusters, CTCs, Normal Cells, and Vertical Cells. Based on the features, cells are classified using a two-step Random Forest and their accuracy is evaluated. Furthermore, the effectiveness of the proposed method is demonstrated by comparing it with conventional methods.

We propose a method of designing a FRIT(Fictious Reference Iterative Tuning)-based control system for a microwave plasma CVD(Chemical Vapor Deposition) apparatus used in the preparation of diamond thin films. Specifically, we propose a control system that adjusts the flow rate ratio of CH₄ and H₂, the material gases, using the spectral emission intensity ratio of the plasma, which affects the film quality, as the controlled variable. We show how to design an experimental apparatus including actuators and sensors for this purpose. The FRIT method, a well-known data-driven control method, is used to design the controller gain from a set of experimental data without modeling complex objectives such as plasma behavior. Through control experiments of the emission intensity ratio, it is shown that the controller gain can be tuned by using a reference model with different time constants to adjust the performance of the response of the control system.

In this paper, general-purpose uniform temperature control for heating process systems is described. The uniform temperature control function should be implemented to a local PID temperature controller. The controller is supplied to the heating process equipment manufacturer. In composition of the business, the adequacy of the relations between companies of the control technology is an important issue.

Self-repairing PID control has been proposed as an active fault-tolerant control system that tolerates component malfunctions, maintaining desired performance. In this method, an auxiliary signal with the same sign as the control quantity is added to the integrator to cause minute oscillations in the control quantity, and a fault is diagnosed by detecting the bias in the integral value due to the occurrence of a fault. This method has the tradeoff that a larger auxiliary signal reduces fault detection time but negatively affects control performance. This paper proposes the auxiliary signal whose value is determined by the function of the control quantity to improve the performance of the self-repairing PID control. Simulation results show that the function of the auxiliary signal effectively reduces detection time and improves control performance. Additionally, the performance of the system is evaluated while varying the design parameters, and some insight into the design guidelines for the function is found.

Adaptive output feedback control based on output feedback exponential passivity (OFEP) has a simple structure and strong robustness in regard to disturbances and system uncertainties. However, it is difficult for most nonlinear systems to satisfy the conditions of OFEP. Thus, the introduction of a suitable parallel feedforward compensator (PFC) to construct an OFEP augmented system with the controlled system has been used, but the control output cannot achieve perfect tracking because of the output of the introduced PFC. As a solution, introducing a feedforward (FF) input to build a 2 degree of freedom (2-DOF) is a simple and effective way to solve this problem. In this paper, we propose the design schemes for suitable PFC and FF input of nonlinear systems via the use of neural networks (NN) respectively. And besides, to cope with possibly present input disturbances, we also provide a method to achieve disturbance compensation based on NN to reduce their interference. Finally, the effectiveness of all proposed design schemes is confirmed through numerical simulations.

With the arrival of the Society 5.0 era, STEAM education is attracting attention, and the Sixth Science, Technology, and Innovation Basic Plan (2021) includes STEAM education for the first time as a science and technology policy, and efforts are underway in school education. In addition, programming education in elementary schools has been made compulsory from 2020, and this has led to the enhancement of programming education in general education. However, there are few practices that incorporate measurement and control technology as part of programming education in elementary schools. In this study, we developed measurement and control technology learning materials for elementary schools that can be practiced as STEAM education. Using kinetic art as a learning material, we proposed a learning content of making a peacock that moves by measurement and control and conducted a class for 6th grade elementary school students. After the class, we conducted a questionnaire survey of the students and their homeroom teachers to verify the usefulness of the learning materials and the class.

In order to realize the "Society 5.0" proposed by the Cabinet Office, human resource development to create new value is attracting attention. In addition, the industrial world is promoting "model-based development (MBD)," which enables efficient development through simulations using mathematical models. In this study, we develop and implement a curriculum to promote model-based development (MBD) thinking at the high school level and test its educational effects. Specifically, the curriculum is to design an advanced emergency braking system through simulation, and to verify the simulation results by using the operation of automotive learning materials.

In this research, we propose a method to adaptively generate AUV trajectories using a neural controller that learns by applying a genetic algorithm. The method uses two neuro-controllers. One is learned in advance as a controller. The other is learning at the same time as redesigning a new trajectory when the AUV deviates from the planned trajectory during navigation. The two neural controllers are evolutionarily learned using genetic algorithms. In the navigation simulation, the multi-point search problem of AUV under the tidal current environment was taken up. Numerical experiments have shown the effectiveness of the proposed method.

The present study proposes a design methodology for blade pitch angle control of offshore wind turbines. A wind farm is formulated as a multi-agent system, and an event-driven distributed control system is designed to reduce the number of data communications.

In general, it is known that highly accurate weight measurement is not possible when the measuring instrument or the object to be measured is shaking. The weight is a DC component, and the AC component is noise. Hilbert transformer can estimate the frequency corresponding to the noise, but contains error. In this paper, we show a method for accurately measuring the weight of a liquid by correcting the error caused by the Hilbert transformer.

In order to analyze check the class individually by making our rounds in the classroom quantitatively, it is necessary to develop a system that replaces the existing analysis method of visual video recording. The purpose of this study is to develop a system that records the trajectory of a teacher using a two-dimensional sensor. The position of a teather is measured from the point cloud data output by the sensor, and after background subtraction, the point cloud data is calculated by grouping the student's point clouds. Experiments were conducted to examine the validity of the position measurement method used.

1985年のオーストリアでのフィラハ会議にて，地球温暖化（global warming）が世界共通の課題と認識されてから40年近くが経とうとしていますが，世界各地での山火事や洪水，国内においても異常な猛暑や線状降水帯の発生による災害級の大雨等，連日のように目を覆いたくなるような

Efforts to achieve carbon neutrality are accelerating in order to solve global warming, a worldwide environmental problem. Efficient use of electric energy requires improved performance, higher functionality, and higher quality of power semiconductor devices, which are key components. In recent years, as the performance of silicon (Si)-based power devices is approaching the theoretical limit, there is a strong need to not only improve performance by optimizing the structure, but also to fully exploit the performance of the devices through device-usage techniques. Silicon carbide (SiC) and gallium nitride (GaN) devices, which are wide bandgap semiconductors, are already in the market, but in addition to further improvement of their characteristics, improvement of their reliability and durability is also required. In this review, we will introduce the technical trends of power semiconductor devices, focusing on the latest papers.

Diamond has excellent physical properties as high breakdown electric field, high thermal conductivity, and high bulk carrier mobility. From these points of view, diamond is expected as an appropriate material which can be applied to next-generation high-power devices. However, the impurity levels of p- and n-type diamond are very deep. Therefore, carrier concentration generated by impurities is very low at room temperature. In order to overcome this issue, heavily impurity-doped layers, which shows hopping-conduction, can be applied in diamond devices. In this paper, Schottky-pn diode with extremely high forward current density is introduced as a typical diamond diode using hopping conduction layers. The recent progresses of diamond switching devices, such as MOSFET, are also introduced.

Surge voltage at IGBT turn-off switching was analyzed with dependence on cell design parameters. Although drift layer thinning is effective to improve trade-off characteristics between turn-off loss E_off and on-state voltage V_on, voltage surge is induced due to quick expanding depletion layer. Therefore, the surge voltage V_surge has also trade-off relationship with the V_on at the same E_off condition. The origin of voltage surge was analyzed using TCAD simulation, and total amount of remained hole in the drift layer during turn-off switching is a key factor for the V_surge. Narrow mesa structure and thick buffer layer are effective for improvement of trade-off characteristics between V_surge and V_on.

The mechanical short-circuit failure mode of 1.2 kV class SiC MOSFETs was investigated by experiment and electro-thermal-mechanical technology computer-aided design (TCAD) simulation. In this work, the details of the simulation procedure are described. TCAD simulations provide a powerful means to address the mechanical failure of such devices, and indicates that high-level stress concentrates at the gate dielectric layer because of the large thermal expansion coefficient of the source metal. Further, a low mechanical stress design for short-circuit in SiC MOSFETs is proposed by using low thermal expansion metal such as copper as the source electrodes.

Insulated gate bipolar transistors (IGBTs) are now widely used in a variety of products ranging from home appliances to power conversion equipment. This paper introduces multi-gate control techniques for IGBTs with the aim of further reducing their total power loss. Results of experiments using prototype devices confirmed that multi-gate control achieves reductions in turn-off loss (E_off), turn-on loss (E_on) and reverse recovery loss (E_rr) by 27%, 50% and 32%, respectively, compared with the conventional single-gate controlled devices.

This paper presents the negative drain input measurements of fully isolated nLDMOS which is fabricated by a low-cost process without any additional epitaxial growth. The critical drain current which causes the parasitic PNP activation is proposed as the index of the negative drain input withstand capability. The device size dependence measurements show that the negative drain input withstand capability decreases as the internal LDMOS area increases which is surrounded by the n-type isolation layer electrode. And, the bias application measurements to n-type isolation layer show that the trade-off relation between the anomalous substrate leakage and the parasitic PNP activation; that is, the higher applied bias suppresses the parasitic PNP activation but makes the anomalous substrate leakage larger.

A concept on High-conductivity IGBT with Time And Space Control of stored carrier (TASC HiGT) for pursuit of low-loss is proposed. The function is divided into high-conductivity chip (Hc) and high-speed chip (Hs) controlled by Dual-gate so as to exhibit the optimum stored carrier density for conduction and switching operation. By obtaining controllability of the stored carrier density over the drift, a ‒47% trade-off improvement on 6.5 kV rated was demonstrated. The proposed concept can generate further ‒34% inverter loss effect keeping low-cost Si material merit.

In this paper, a newly developed 3^rd Gen. RC-IGBT is presented. To reduce the power loss and improve the heat dissipation of RC-IGBTs, we introduced four approaches, controlling Gate-capacitances, thin wafer technique, high-density diode arrangement, and low carrier injection anode and cathode. The superior performance of the 3^rd Gen. RC-IGBT is very promising for the carbon neutral.

We fabricated high forward and low leakage current trench MOS-type Schottky barrier diodes (MOSSBDs) on a 12-μm-thick epitaxial layer grown by halide vapor phase epitaxy on β-Ga₂O₃(001) substrate. The MOSSBDs, measuring 1.7×1.7 mm², exhibited a forward current of 2 A (70 A/cm²) at 2 V forward voltage and a leakage current of 2×10^-8 A/cm² at -1.2 kV reverse voltage (on/off current ratio of > 10⁹) with an ideality factor of 1.05 and wafer-level specific on-resistance of 17.1 mΩ·cm². These results indicate that the MOSSBD is promising for practical high-voltage (AC 400 V) power applications.

In this work, we report a newly developed analysis method to investigate the mechanism of turn-on voltage tail of IGBT, aiming at the reduction of turn-on loss (E_on). This analysis method visualizes how gate-collector capacitance (C_GC) increases by monitoring the potential change at the interface of the gate oxide and silicon around the gate trench during the turn-on process. It was found that, in the turn-on transient, initially, the collector-emitter voltage (V_CE) decrease, and C_GC increase drastically when the potential around the gate trench falls below the gate-emitter voltage (V_GE) and that the sudden increase of C_GC causes the turn-on voltage tail. Detailed TCAD analysis reveals that the electric potential around the gate trench is lowered by the dummy trench and causes the sudden increase of C_GC. To verify that the dummy trench is origin of the turn-on voltage tail, full gate trench IGBT with no dummy trench is studied. In the full gate trench IGBT, neither sudden increase of C_GC nor turn-on voltage tail are observed.

Back-gate-controlled IGBT (BC-IGBT) which uses a manufacturable double side lithography is experimentally demonstrated. The impact of IGBT scaling on BC-IGBT is evaluated in detail. A back gate control (BC-) mode in scaled IGBT shows better performance improvement compared with the BC-mode in non-scaled IGBT. Superior performance of BC-IGBT by optimizing both front side and back side design has been realized. Back gate control with scaled IGBT will provide a new technological option for expanding the frequency/voltage range of Si power devices, and become an attractive candidate for next generation high performance IGBTs.

We propose a new edge termination technique for high-voltage vertical GaN power devices using boron-ion implantation into a p-GaN epitaxial layer. The boron-implanted layer after annealing at 800°C results in a half-conductive p-type layer that is applicable to junction termination extension (JTE). The maximum breakdown voltage of 1400 V was near an ideal value for the donor concentration in the drift layer. High avalanche current immunity was observed for the optimal structure. In wafer-scale fabrication of the p-n diodes on a 4-inch free-standing GaN substrate, good reverse characteristics are observed across the wafer. The proposed JTE technique has high potential for wafer-scale fabrication of robust and highly efficient vertical GaN devices.

We compared static and dynamic characteristics of a 1.2 kV-Class SJ-MOSFET fabricated by the multi-epitaxial growth (ME) method and the trench-filling epitaxial growth (TFE) method, which is expected to have a lower process cost. Comparably low specific on-resistance and switching losses were observed for the SJ-MOSFETs fabricated by both methods.

This paper proposes a low on-resistance and low reverse recovery charge Power MOSFET Technologies for large-capacity communication systems. We have developed 150 V class power metal-oxide-semiconductor field-effect transistor (MOSFET) products featuring low on-resistance, gate charge, and reverse recovery charge characteristics by deep trench field plate technology and lifetime control technique. On-resistance reduced by 44% and reverse recovery charge reduced by 74% compared to the previous generation product. The device achieved peak power conversion efficiency of 94.6%.

We propose a sound source separation method to extract target sounds from multiple sounds observed at binaural microphones worn by a user. The proposed method estimates the respective sound source position from the slope and variance of the spectral phase difference of two signals observed at the binaural microphones. Based on the estimated sound source positions, the proposed method separates all the sounds and extracts only the sounds that arrived from the target area specified by the user. Simulation and real-environmental experiment results show that the proposed method efficiently gives the sounds whose sources exist in the target area.

In the aging society, especially those with eye diseases, often forget to take medicine or take wrong medicine. In the new drug approval process, objective and accurate medication records have become very important. To improve the reliability of medication recording, we have been developing a cup-shaped medication recording device by taking intraoral pictures. This device has a camera with a wireless transmitter and a rechargeable battery in the cup. To recognize oral medicine accurately, we used AI image recognition technology. We selected four representative drugs (500 images for each) for testing, and the recognition accuracy can reach more than 90%. While ensuring the recognition accuracy, we achieved real-time detection, which can reach the detection speed of 16 frames per second on smartphones. In addition, an Android application based on AI image recognition has been developed that is easier to use. The medication management application was equipped with image acquisition, image recognition, medication recording, forgetting of medication, and caution for wrong medication.

We investigated the possibility of eye blink detection for one person using acceleration sensors fixed on a diving mask. It was found that the two accelerations measured on the upper and lower parts of the mask skirt were changed inversely during voluntary blinking. Then, we calculated the correlation coefficient of these accelerations, and the eye blinks could be detected with over 80% accuracy.