In this paper, we study the design of a simple adaptive control system for the leader-following problem of discrete-time SISO multi-agent systems with uncertain heterogeneous dynamics. The multi-agent systems are assumed to have directly designable agent, called a leader, and non-leader agents, called followers. Each follower is given a reference model with the same dynamics as the leader and a parallel feedforward compensator (PFC) to ensure that the system has ASPR. The control system is designed so that the output of the augmented system of the follower with PFC follows the output of the leader, and the tracking error is guaranteed to be bounded by using Lyapunov functions. We also present the numerical simulation to demonstrate the effectiveness of the proposed method.

This paper presents a PID gain tuning method based on model matching to desired reference models for closed-loop systems with PID controllers. The features of the proposed method are as follows. Even if the plant transfer function model is unknown, if its frequency characteristics can be obtained, the PID gain can be tuned, and the stability margin and time responses of the closed-loop system after tuning can be estimated. Furthermore, by newly introducing an approximate model matching error, the least squares method can be applied to obtain the PID gain, and an appropriate reference model that satisfies the desired stability margin can be efficiently searched for. The usefulness of the proposed method is demonstrated through numerical examples.

In control systems using network communication or digital sensors, the quantization of input and output signals can significantly impact control performance. Therefore, it is essential to incorporate signal quantification into the control system design to prevent performance degradation. This study proposes a controller design method for systems in which both inputs and outputs are quantized. The proposed method derives a controller that minimizes the upper bound of the error between a plant output and a reference model output. In contrast to previous studies, our proposed method avoids solving matrix inequalities numerically and provides an analytical design approach.

The uniform temperature control for heating process systems in semiconductor equipment is implemented to a temperature controller which is supplied to the semiconductor equipment manufacturer. In composition of the business, the adequacy of the relations between companies of the control technology is an important issue. In this paper, comparison of logical approach and data-driven approach of the uniform temperature control is described.

Our previous work has developed a self-repairing PID control (SRPIDC) for sensor failures. This detection method uses the I-controller (integrator) of the PID control and an auxiliary signal to detect failures of a stuck type. Since the auxiliary signal varies in the same direction (sign) as the feedback signal from the sensor, the magnitude of the integrator output increases when the feedback signal gets stuck. Such an unstable behavior is found by a threshold. However, when noise is mixed in the feedback signal, this detection function does not work well because the sign varies. In this paper, we quantitatively analyze the degradation of the detection performance of the SRPIDC due to noise and propose two methods using an on-off type dead zone and a piecewise constant switching signal, that can make the sign of the shaped signal fixed on periods to detect sensor failures in the presence of noise.

The COVID-19 pandemic in 2020 has characteristics such as asymptomatic infections and reinfection, and the characteristics have not seen in previous SARS. An SIIR model was proposed as the mathematical model that considered such characteristics, but the control strategies, for example, vaccination and quarantine, was not considered in the model. Therefore, in this paper, we propose an SIIQR model that employs vaccination and quarantine as control inputs to the SIIR model. We also obtain optimal control using Pontryagin's maximum principle in order to suppress infectious people as well as minimizing the costs of vaccination and quarantine. Finally, the validity of the proposed method is confirmed through numerical simulation.

This paper focuses on the development of learning methods for achieving effective collaborative transportation by multiple robots in a warehouse environment. In large-scale and complex environments, it is necessary for agents to undergo numerous iterations of learning, such as reinforcement learning, to make appropriate behavioral choices. Traditional multi-agent methods like MADDPG (Multi-Agent Deep Deterministic Policy Gradient) and QMIX face the issue of requiring extensive computation time for environmental exploration. Therefore, this paper proposes a two-stage learning procedure that separates overall optimization, including the formulation of general task execution procedures, from individual optimization based on local situation assessments. Additionally, the effectiveness of the proposed method is demonstrated through simulation system analysis adapted to the target environment.

The planned operation of work equipment leads to higher productivity at construction sites. Predicting breakdowns and performing maintenance in advance are important for the planned operation of work equipment. If the load applied to a machine in operation is known, breakdowns can be predicted based on the condition of the load. However, the load applied to construction machinery, such as hydraulic excavators, varies depending on the work situation and environment. The load applied during excavation operation varies depending on the excavation conditions, such as the excavation target, soil volume, and motion conditions. To estimate the load applied to attachments during excavation, it is necessary to use a load estimation model according to the excavation conditions. Excavation operations are affected by the external environment, so multiple load estimation models are needed. Therefore, it is not practical to create models in advance and switch between models. This study proposes a database-driven load estimation modeling method for excavation operations. The proposed method constructs a database containing operational data and model parameters. Then, the parameters are identified from data similar to the requirement point, and a local model is created. The effectiveness of the proposed method is verified through experiments using a radio-controlled excavator.

Digital differentiator is digital circuit for differentiation of digital signal. It is often used in a wide range of fields such as sound processing, image processing, measurement and control. In this work, we propose the design of the sparse integer order differentiators which contain zero coefficients. We show the fast search algorithm for obtaining the solution of sparse integer order differentiators. Design examples show that the fast search algorithm can often find the optimal solution which is obtained by the tree structure. Thus, the effectiveness of the fast algorithm can be demonstrated.

In recent years, the development of automated driving technology has been actively pursued in line with changes in the automotive industry. In the real-world environment where automated driving is performed, there are many uncertainties such as changes in road surface conditions due to weather and terrain, changes in vehicle weight due to the number of passengers and luggage, and changes in vehicle speed. These uncertainties deteriorate the performance of the control system and significantly affect the driver's comfort and safety. In this paper, we propose an adaptive output feedback control based on almost strictly positive real (ASPR) for a path-following control system for four-wheeled vehicles that can cope with uncertainty in vehicle characteristics. In order to maintain robustness against system uncertainty, a parallel feedforward compensator (PFC) will be introduced into the control system to keep the augmented system ASPR when the vehicle's ASPR performance cannot be guaranteed due to changes in parameters such as speed. The effectiveness of the proposed method is confirmed through numerical simulations.

To investigate the improvement in skill of rookie through training and the difference in skill between rookie and expert, the steering wheel steering angle and bank angle during straight narrow driving were measured. After training, rookie increased their steering effort, steering range, and appropriate corrective steering to the tilt of the vehicle. Expert corrected the vehicle to the left and made faster corrective maneuvers to vehicle lean.

This paper considers an event-triggered LQ control of a hung-type ball-and-beam (HTBB) system via wireless communication. In this HTBB system, the beam angle is controlled by winding the wire. Unfortunately, due to gravity, the beam's angular acceleration cannot be greater than its free-fall acceleration. In addition, discontinuous changes in the beam angle sometimes cause centrifugal force. To solve this problem, we introduce two integrators in the front of the HTBB system. It makes it possible to render the angular acceleration continuous and small in the sense of LQ optimization. Although the event-triggering mechanism (ETM) reduces the number of communications, it may cause unnecessary updates of the control input. Therefore, this paper proposes a method to avoid this issue by introducing a dead-zone to the ETM. Several experiments are explored to confirm the effectiveness of the proposed method.

This paper considers the anomaly detection of mobile robots in the consensus control, whose objective is to that states of all agents in a multi-agents system converge to common state value by exchanging information over a network. In the presented scheme, one class support vector machine, that is one of unsupervised learning approaches, is used and multiple classifiers are designed to deal with two cases: i) a failed robot does not move from the initial position and ii) a failed robot continues to move in a certain direction. Simulation and experiments with omnidirectional mobile robots show that anomaly detection accuracy is improved by using both two classifiers.

In this paper, a novel method for solving multiple sound source localization problem using two microphone system. In this problem, it is assumed that the number of sound sources is more than 2 and is unknown in advance. That is, this problem is one of underdetermined problems having unknown number of measurement targets. For solving such a problem, a method based on swarm intelligence is applied, in which estimation of the number of sound sources and multiple source directions in the same framework. The effectiveness of the method is shown from several experimental results in the actual room.

To improve the accuracy of the sound image localization technique using bone conduction headphones and head-related transfer functions, this paper proposes a preprocessing method using the difference of transfer characteristics between from the entrance of ear canal to cochlea and from transducer to cochlea. The estimation of the differential characteristic is performed using Distortion Product OtoAcoustic Emission (DPOAE) and the preprocessing is realized finite impulse response (FIR) digital filters. In subjective experiments on sound image localization, it was confirmed that the proposed method improves the localization accuracy compared to the case without preprocessing.

かつて1980年代に，日本の半導体は全世界のシェア50％を占め，世界を席巻しておりました。しかしながら，日本の半導体は90年代以降その存在感を急激に失い，近年のマーケットシェアが10％に低下するに及び，日本政府は強い危機感を持って半導体デジタル戦略の再構築に踏み切って

This article reviews the development of Si-based thermoelectric device conducted by the author’s group. A series of integrated thermoelectric devices of silicon-nanowire thermoelements were fabricated with a matured and standard Si-CMOS process, demonstrating scaling rules for miniaturization and large-scale integration on thermoelectric performance. The thermoelectric device exhibited a high specific power generation capacity of 0.9 µW/cm²K². The optimum device layout to maximize the power generation capacity will be discussed in detail.

2-dimentional materials are now intensively studied with the aim of the application to advanced CMOS logic transistors for 1 nm node and beyond. Here, the reason why the 2-dimntional materials are expected and the present status of the research of 2-dimntional materials transistors are reviewed. Challenges and prospect of this technology are also discussed.

A method to reduce the high coercive field (E_c) of ferroelectric (FE) Al_0.7Sc_0.3N films was proposed. By proper oxygen incorporation to form Al_0.7Sc_0.3O_0.2N_0.8, E_c reduction from 3.8 to 2.8 MV/cm was achieved while maintaining a high remnant polarization (P_r) of over 90 µC/cm². The E_c reduction is presumably caused by the O and Al atom complex vacancies; different from the Sc-atom induced distortion to decrease the ratio of wurtzite lattice constant (c/a). A dielectric constant increase from 17 to 30 supports the E_c reduction. By designing the oxygen depth profile in the film, FE property was obtained with a thickness down to 7 nm. Even the E_c increase with thinner thickness is still presented, a ferroelectric switching was confirmed at a voltage below 5 V.

For scalable semiconductor-based quantum computer, integration of qubit, electrometer and readout circuit on CMOS platform needs to be explored. Here, for qubit readout, the use of charge-based capacitance measurement (CBCM) is proposed to detect the atto-farad change in the gate input capacitance of single-electron transistor (SET) electrometer. Mixed SET-CMOS circuit simulation revealed that a few hundred microvolts signal appeared at the integration capacitance at 1MHz data rate and the gate capacitance profile with respect to the gate voltage was successfully reproduced.

This paper describes a new ultra-low power and ultra-low frequency oscillator using the GIDL current used in edge AI devices for long-term and intermittent monitoring applications. Various characteristics of GIDL current were measured and compared with the subthreshold current. An ultra-low frequency GIDL ring oscillator was designed and evaluated based on the measurement results. As a result, ultra-low frequency oscillation from 0.94 to 5.1 Hz was successfully achieved. An extremely small circuit area was achieved because integrated resistors and capacitors were not used. From the power consumption analysis, a new low-power design was conducted from the viewpoints of "shortening the signal transition time of the buffer circuit" and "reducing the short-circuit current of the buffer circuit," and ultra-low power operation of 71.6 pW was successfully achieved in circuit simulation. GIDL ring oscillator is a promising candidate for an ultra-low power and ultra-low frequency oscillator with high environmental stability.

In this study, we created a skin-parallel type transcutaneous transmission system for a ventricular assist device using plate-shaped solenoid as the transmitting coil. We compared the biological effects and radiated magnetic fields of six models in which two receiving coils were placed in series and in parallel. The results showed that the power transmission efficiency was around 90% for all models, and there was no major difference between the two layouts. It was also found that by placing two receiving coils in parallel and separating them by 10 mm (Type D), was possible to suppress the electric field strength within the body E_in by more than 10 V/m. However, the limit value was exceeded by placing two receiving coils in series and separating them by 0.5 mm. In addition, the local SAR for Type D was 1.2595 W/kg, which was below the limit value.

The inspection of ground lines on transmission towers is necessary to maintain stable power transmission. We already developed a UAV which automatically acquires images of the ground line for inspections. In this paper, we propose a method for detecting defect locations of the ground line from the images acquired by the UAV. In order to identify the defect locations, it is necessary to extract a ground line region from the image in advance. The proposed method finds the ground line from multiple line segments in the image by using edge intensity of the image. The identification of defect locations is performed by a deep neural network (DNN) for each divided image of a ground line region. As a result of the experiment, a success rate of 0.978 was obtained for the extraction of a ground line from UAV images. In case of using the GoogLeNet as the DNN, an accuracy rate of 0.975 was obtained for defect detection.