We developed full color LED display with color sensor function by using LED to an optical sensor. This display has a new function that you can draw pictures and letters by an irradiating light. The drawn color is selected in 3 patterns by a color of the light irradiation. If red light is irradiated, red LED in the display turns on, if blue light is irradiated, green LED turns on and if UV light is irradiated, blue LED turns on. In this paper, a measurement method and its results of wavelength sensitivity characteristics for a full color dot matrix LED, an operation principle and a structure of the circuit in the display are discussed.

In this research, we examined whether we could estimate the position of the transmission source by using the Bluetooth beacon with Eddystone sent by a smartphone. In our experiment, free-space attenuation and soil attenuation of the beacon were measured. From the experimental results, when field intensity more than -80 dB was obtained, the source was within 5 m in the free space and within 0.5 m in soil-depth. Therefore, this method showed the possibility that was effective for victims search.

For achieving the encoding and the decoding of high-resolution digital video streams, high efficiency video coding (HEVC) is implemented to many devices. HEVC can encode input video with a high compression ratio by adopting MC (motion compensation) -DCT method. Additionally, HEVC can control the video sequence quality by a quantization parameter (QP). However, there is high possibility that redundant bits are generated by MC-DCT method which calculate the rate-distortion cost by the prediction method, variable block size, QP, etc. In this paper, the generated bits are adjusted by QP value based on saliency-detection map. The simulation result show that the proposed algorithm do not affect the subjective evaluation and achieve 11.02% average reduction of bit-rate.

Analysis precision of buck ratio and power efficiency in buck DC-DC converter circuit is improved by proposing adaptive equivalent circuit of flywheel diode in the circuit. In conventional analysis, the flywheel diode was replaced with single resistor of constant resistance independent of change in average diode current. Therefore its analysis precision was limited caused by large difference in voltage-current characteristics of the replaced resistor between that of the actual flywheel diode. We proposed replacing the flywheel diode with series circuit of resistor and voltage source changed their resistance and voltage according to change in average diode current. In analysis results, error in theoretical value compared with measured value of the buck ratio can be reduced from 23.7% to 6.61% in a conventional buck DC-DC converter circuit, and from 20.4% to 9.24% in tapped-inductor low buck ratio buck DC-DC converter circuit. In the analysis of power efficiency, error in theoretical value compared with measured value can be reduced from 22.5% to 5.45% in the conventional circuit, and from 25.4% to 4.76% in the low buck ratio circuit.

A new radio system that deterministically uses reflection waves is proposed. This system transfers data on the rotating polarization wave (RPW) which rotates the polarization at enough lower frequency than that of propagation to digitally control it. The receiver discriminating a deterministic reflection wave of the RPW by selecting the polarization that the transmitter uses achieves good non-line-of-sight communication quality.

Maintenance support systems (MSS) monitoring conditions of facilities are indispensable for condition-based maintenance (CBM). A sensor network is used to acquire maintenance data in the MSS, and installation of the sensor network should be simple and easy to execute maintenance tasks in a cost-effective way. Therefore, this report sets a proposal of an autonomous configuration function of the systems that makes the installation simple and easy. This function is composed of wireless sensor networks, plug-and-play function configuring software modules, and cooperative procedure between them. This paper also shows evaluations of the proposed function that is feasible for practical maintenance work. The evaluations have been carried out using a prototype system implementing the elements of the proposed function.

This paper proposes a robust compensator design for sinusoidal inverter based on the design methodology named Dual Model Matching and numerical optimization. This proposed method can simply design a controller which ensures high speed tracking with low distortion and zero steady-state errors for sinusoidal voltage reference. The software-in-the-loop simulations are presented to support the advantages of the design and to demonstrate the system performance.

The distance to a target is basic information in many engineering fields. As is generally known, the distance is typically estimated using time-of-flight (TOF) between transmitted and reflected impulses. However, it cannot be estimated in a close range where the reflected waves are buried in the transmitted wave. For distance estimation of close range targets using acoustic signals, we have proposed a method based on the phase interference between transmitted and reflected waves. In addition, by using cross-spectral method, we have proposed a method to remove the effect of the measuring system which is composed of loudspeaker, microphone, amplifiers and so on. Moreover, this method can also utilize environmental sound such as music and speech with arbitrary frequency spectrum as a transmitted sound besides a band-limited impulse sound. This paper describes a prototype system realizing this method by implementing it into a microcomputer system and confirms the result of experiment in actual sound field using music sound source.

Finger-spelled sign language recognition is one of hand sign recognition tasks and one of most challenging ones in some research area such as computer vision and human computer interaction. In this research, we especially focus on JFSL (Japanese Finger-spelled Sign Language). To realize a practical system for automatic finger-spelled sign recognition, we need to analyze the sequentially represented signs in real-time with some complex processes such as the recognition of signs one by one, word recognition, morphological and sentence analyses, etc. Although many finger-spelled sign language recognition researches have already been existed, they mainly focus on the one-by-one sign recognition and discuss the analysis of sequentially represented signs insufficiently. Therefore, compared with them, we focus on the spotting system of sequentially represented signs to move the next step of one-by-one sign recognition. In this paper, we propose a new spotting techniques for JFSL with SVM and rule-based approaches, which are based on moving direction of hand and finger directions before and after hand movement. From the experimental results, we achieved the 81% of average division rate for division of sequentially represented signs into one-by-one signs.

In recent years, analytical method in which the length of time-window for time-localization can be established optionally was shown by the authors. If you decide the time-window width which is in inverse proportion to the frequency, The function becomes the same function as the elastic function of the mother wavelet of wavelet transform, Effective analysis is carried out from low frequency to high frequency. In this analytical method, I use a signal detection unit to detect a signal. Frequency to detect a signal is set in the signal detection unit and I call it established frequency. In this analytical method, a lot of signal detection units which have different established frequency are used. First of all, all signal detection units analyze a signal as an analysis object. This is the analysis of the first stage. The wave which appears to each signal detection unit provided by the analysis of the first stage is analyzed by the signal detection unit. This is the second stage analysis. I repeat this, and multi-stage analysis is carried out. Until now, it was shown that the displacement in the time axes direction occurred at each signal detection unit in each stage analysis in the multi-stage analysis. In this paper, it is showed that new phase to produce at each signal detection unit in one analysis in the multi-stage analysis is the same as the displacement of time axes direction. Until now, there were amplitude characteristic and a phase characteristic as the condition to get a reconstruction wave. A phase characteristic disappears by revising the displacement at time axes direction and shows that it becomes only the amplitude characteristic. In addition, it shows that a reconstruction wave composes a signal waveform with high precision in a numerical analysis example using the Japanese vocal sound.

In pronunciation training, we learn to properly control properly our articulator for correct pronunciation. Therefore, the pronunciation training system needs to build effective visual support for the trainee. In this report, we propose a pronunciation training system that constructs visual support based on formants, and examine its effectiveness. From the system verification experiments, we identify the problems caused by the fact that the trainee has not recognized the tongue shape. We focus on the protrusion of the lips and investigate the relationship between point of articulation and acoustic features. As a result, it was found that the vocal tract area function, which is an acoustic feature, has a stronger relationship with the point of articulation than the formant. Moreover, we believe that this result can cope with deviation of the point of articulation due to protrusion of the lips and can be utilized for pronunciation instructions.

This paper proposes a new approach to solve Chance Constrained Problems (CCPs) efficiently. Specifically, the probabilistic constraint in CCP can be evaluated directly if the Cumulative Distribution Function (CDF) of uncertain function value is known. Therefore, the CDF is approximated by using Weighted Empirical CDF (W_ECDF). Then a powerful evolutionary algorithm, namely Differential Evolution, combined with W_ECDF is used to solve CCP. In order to demonstrate the performance of the proposed method, it is applied to flood control planning problems.

In recent years, presentation with computer is important as a social skill to transfer knowledge and information, but presentation practice is depending on the know-how, such as instructor's experiences or textbooks. We have developed a presentation practice support system, named PRESENCE, to improve beginners' demonstration of presentation. The PRESENCE has the real-time voice feedback function to recover undesirable states. As the basic skills of presentation to support beginners, the voice feedback function has been implemented to check body expression such as “the vertical direction of face”, “the horizontal direction of body” and voice expression as “the voice volume”. The effects of PRESENCE to beginners' presentation were investigated by comparing those without the feedback function. The results showed that PRESENCE significantly improved the voice expression, and significantly changed the movement of the body expression.

In semiconductor device manufacturing process, it is necessary to measure overlay error between integrated layers. As semiconductor process design rules continue to shrink, small overlay error comes to lead to the fatal fault of the device electrical property. As a result, the needs for the dense overlay measurement in the chip with the use of circuit pattern increase. We propose an automatic extraction technique of the evaluated points (EPs) that are suitable for the overlay measurement by scanning electron microscopy (SEM) using design data of the device pattern layout. The proposed algorithm extracts all measureable patterns by evaluation of the positional relationship between a segment pair on the upper and lower layers by plural selection indices. Simulation was performed to estimate the overlay error distribution within the shot area using the EPs automatically extracted by the proposed method. The standard deviation of estimation errors in the x- and y-directions were 0.06 nm and 0.10 nm (3σ), respectively. It was confirmed that distortion in the exposure apparatus can be estimated automatically and highly accurately.

We develop a new optimization method for the daily water supply pump operation scheduling using the branch and bound method. In the method, we define a search tree as a trinary tree in which the nodes are the number of operating pumps at each step and edges are the changes in the number of operating pumps between the present step and the next step. We also define pruning criteria of the search tree to raise search efficiency. We apply the method to a benchmark of the daily water supply pump operation scheduling, and verify the effectiveness of the method.