In this paper, we introduce a real nursing sensor dataset which includes labeled dataset for supervised machine learning and the big data combined with patient medical records and sensors attempted for 2 years, and also describe a method for recognizing activities for a whole day utilizing prior knowledge about the activity segments in a day and utilizing importance sampling and Bayesian estimation, based on our paper at UbiComp2015 [13]. Moreover, we demonstrate data mining by applying our method to the bigger data with additional hospital data. Our method of recognizing a whole day activities outperformed the traditional method without using the prior knowledge. Moreover, the method significantly reduces the duration errors of activity segments. We also demonstrate a data mining applying our method to bigger data in a hospital, and show several results about the correlations with nurse profiles and patients status using Random Forest regression.

View full abstract

In this article, we propose a design method of a digital filter that suppresses self-interference arising at a single-frequency full-duplex wireless relay, and also compensates frequency selective fading from a source station to the relay station. The signals and the wireless relaying system can be modeled as random processes and a discrete-time linear time-invariant system respectively. We formulate the filter design problem based on the minimum-mean-square error (MMSE) criterion. It is shown that the filter design problem is equal to an H² optimal control problem. The effectiveness of the proposed method is shown by simulations.

View full abstract

This paper investigates nonlinear stochastic optimal control problems under non-Gaussian white noise, which is represented by so-called Lévy process. We show that the path integral approach combined with policy iteration gives a numerical solution to the Hamilton–Jacobi–Bellman equation.

View full abstract

The paper is concerned with non-fagile control design via static output feedback controller for fuzzy stochastic systems that are subject to stochastic disturbances. Many theoretical control design methods given in the literature do not take case of any disturbance into the system and variations in control gains, and they assume all the states of the system are available for the feedback control. In the practical situations, however, unexpected disturbances come into the system and malfunction may occur in actuators. Hence, a non-fragile output feedback control design is desired. In this paper, robust stabilization via static output feedback control with uncertain control gain for fuzzy stochastic systems is considered and a control design of robust stabilizing controllers is proposed. The stability analysis of the closed-loop system and controller design via descriptor system approach are given in terms of LMI conditions, which are less conservative than the existing results. An illustrative example is provided to show the effectiveness of the proposed method.

View full abstract

In this paper, a method for controller design for stochastic systems based on optimally controlled stationary states is proposed. The Key idea in this method is invoking the optimally controlled system states to control the system states so as to the system states tend the controlled states close to desired system states. By computational experiments, it is shown that the proposed method is effective to control the nonlinear stochastic systems described by swing equation.

View full abstract

In recent years, the studies on Brain Computer Interface (BCI) to manipulate computer using an electroencephalogram, not through devices such as keyboards and mice, have been actively conducted. The main purpose of such BCIs is to develop communication tools for handicapped users. In this paper, we investigated differences among behaviors of Bereitschaftspotential (BP) which is one of components of movement-related cortical potentials (MRCPs), during the motor imagery of simple movements and complex movements of fingers. As a result, before and after training, some changes, such as integration or deletion, have been observed in the components related to voluntary movements, such as BP, etc. In addition, the inflow and outflow of information between Fpz and Cz have been confirmed to be reversed at intervals of 0.5 sec to 1 sec before each movement onset.

View full abstract

The results of our experiments on categorical perception of different shades of gray are reported. A special color generator was created for conducting the experiments on categorizing a random sequence of colors into two classes, light-gray and dark-gray. The collected data are analyzed based on constructing (i) the asymptotics of the corresponding psychometric functions and (ii) the mean decision time in categorizing a given shade of gray depending on the shade brightness (shade number). Conclusions about plausible mechanisms governing categorical perception, at least for the analyzed system, are drawn.

View full abstract

This paper discusses hemodynamics in cerebral cortex when subjects taking in glucose perform Paced Auditory Serial Addition Test (PASAT). PASAT can give large load to working memory. Hemodynamics of the subject is evaluated using oxidized hemoglobin concentration, HbO, measured by Near-infrared Spectroscopy (NIRS).We apply multivariable estimated models of auto-regressive form with upper triangular estimates. They are given by hierarchical decomposition analysis, which show properties of each stage.

View full abstract

The Electrocardiogram (ECG) is used for diagnosing heart conditions by recording the small electric waves generated during heart activity. Gois et al. have recently proposed a mathematical model to describe heart rhythms considering three-coupled Van der Pol oscillators, and indicated that the heart rhythms of the cardiac diseases can be shown by changing three coupling parameters. In this paper, we propose two methods for simultaneous estimation of the states and the parameters of Gois ECG model. For the joint state-parameter estimation, we constitute the ensemble Kalman filter from this model to detect the cardiac diseases by estimating the coupling parameters as the state variables in Gois’s model. For the dual space-parameter estimation, in addition, we apply the dual ensemble Kalman filter to estimate the coupling parameters. Finally, we compare the estimation results of these filters.

View full abstract

Artificial respirators are widely used in various scenes. Doctors are required to pay scrupulous attention for the use of the artificial respirators. The ideal setting method of artificial respirator is a setting method in consideration of the pulmonary characteristic of the patient. However, we cannot know the pulmonary characteristic of the patient by the measurement of data. Therefore, we must depend on the experience and the intuition of the doctor for the setting of the artificial respirator now. Purpose of this study are to develop a method to estimate the static P-V curve and the pulmonary elastance of the patient and to set a ventilation condition of the artificial respirator. The static P-V curve and the pulmonary elastance expresses the important feature of the lung, and the static P-V curve is a basis for deciding the air-way pressure limit value. In our previous work, we suggested the estimation method of the pulmonary elastance using a recursive parameter estimation. In this study, a parameter estimation of the fuzzy rule’s in estimation technique is improved. The patient data used for parameter estimation are weighted by a conformity degree of the fuzzy variables. Therefore, the data of big conformity degree become important in parameter estimation. In the experiment, a estimation example using real patient data is given to illustrate the superiority of the proposed method.

View full abstract

The number of arteriosclerotic patients has been increasing on account of Japanese Westernization. A diagnostic method using echocardiography is a simple and easy method to diagnose arteriosclerosis. The method measures the intima-media thickness (IMT), which is the thickness of the carotid artery wall. Accordingly, the method employs this thickness as a diagnostic index. When using this method, however, it is difficult to detect arteriosclerosis in its early stages when the artery wall has not yet thickened. Therefore, we propose a method that can be used even in these early stages by employing echocardiogram brightness as a diagnostic index. Because the artery wall hardens in arteriosclerosis progression, the echo brightness shows a large numerical value. In addition, in the possible case of an artery wall exploding, the echo brightness shows a small value because the vascular wall is soft. Nevertheless, we do not know the number of classes in advance. Thus, we must automatically change the number of classes for each image that is obtained.

View full abstract

This paper considers a moving sound source tracking problem and show that the direction of the moving sound source can be sequentially estimated by our frame-wise DOA (Direction of Arrival) estimation approach, which has been originally developed for a fixed sound source by using only mixed signals observed at a pair of microphones. From several simulation results, it has been confirmed that our approach can successfully track the moving sound source in a real time manner under the environments where the signal-tonoise ratio is more than 20dB and the reverberation time is within 200msec.

View full abstract

Acoustic waves are observed as phase vibrations of Rayleigh backscattering in fiber-optic distributed acoustic sensing (DAS). In DAS, spatial derivative of phase represents the amplitude of an acoustic wave. Since measured phases are wrapped in (−π, π], phase unwrapping and differentiation are necessary to detect acoustic waves. In this paper, we propose a phaseunwrapping method in DAS using the Kalman filter and show that the acoustic waves are faithfully detected from noisy phase observation data.

View full abstract

A new class of pseudomeasurements for discrete-time stochastic systems are derived from continuous-time linear stochastic systems with unknown parameters by applying time-discretization and Taylor expansion. Utilizing these pseudomeasurements, we propose new iterative methods which estimate the states of the discrete-time stochastic systems and identify the unknown parameters simultaneously. A numerical example is given to show the effectiveness of the proposed approach.

View full abstract

This paper provides an estimation method of the uncertainty of the pulse transfer function from the covariance matrices of the estimates of the vectorized system matrices in PO-MOESP method, in which Kalman gain and the covariance of the innovations process are consistently estimated by using the previously proposed method by the author. In the conventional methods, the estimates of Kalman gain and the covariance of the innovations process are asymptotically consistent, i.e., the estimates converges in probability to the true value as not only the data length but also the past horizon go to infinity. The contribution of the paper is to provide a consistent estimate of the frequency domain uncertainty for fixed past/future horizons.

View full abstract

A rotary-type unstable inverted pendulum system is modeled by means of a parameter-dependent lineartime-invariant (PD-LTI) system. The pendulum systems operating at upright and downright positions are modeled by a PD-LTI system with a scheduling parameter that is fixed in each operation. The unknown parameters are estimated by taking a grey-box modeling approach to the pendulum system operating at the downright position. The obtained model is verified by designing an H^∞ controller.

View full abstract

This paper is concerned with material interrogation methods for homogeneous dielectric materials using stochastic inversion. Frequency dispersions for homogeneous dielectric materials can be described by the Debye formulation. Propagation in the dielectric medium is assumed to be governed by the time-dependent Maxwell’s equations in one spatial dimension. When a plane wave travelling in the free space hits the dielectric material, the reflective wave from the front edge of material is measurable at the free space. Thus the problem considered here is to quantify uncertainties of the estimated dielectric parameters from the measurements of electric field. A computational method using Hamiltonian Monte Carlo sampling based on nonlinear hierarchical models is tested.

View full abstract

In this paper, an estimation procedure of parameters such as a damping coefficient and a natural frequency in the linear stochastic dynamical system (LSDS) driven by colored noise sequence is proposed based on a methodology applied bi–variate autoregressive (AR) modeling procedure. Ship’s roll motion, in which external disturbances such as waves and winds have strong frequency dependent, is treated as the subject. The verification of the proposed procedure is implemented by using results of onboard experiments that a ship ran in an octagon. Obtained findings are reported.

View full abstract

In this paper we show a data assimilation framework which gives good estimation of the macroscopic parameter in particle systems. To give the fast and stable estimation, we employed the bounded Gaussian uniform mixture (BGUM) type dynamics that is originally introduced in the econophysics field. The result of the numerical experiment implies that BGUM type dynamics enable us to obtain appropriate estimation of macroscopic parameters from macroscopic observations faster than random walk type dynamics that is usually employed. It is also implied that mixing rate between Gaussian and uniform distribution can control the trade-off between fast detection and stability of macroscopic parameters. Those results suggest that the utility of the introduced framework for macroscopic parameter estimation in particle systems.

View full abstract

The author presents a new discretization method of stochastic differential equations (SDEs) without proof. The method is a kind of K-scheme (Kusuoka approximation, Kusuoka-Lyons-Ninomiya-Victoir method). K-scheme is a new framework of the higher order discretization methods for approximating SDEs weakly, proposed by Kusuoka. Ninomiya-Victoir and Ninomiya-Ninomiya methods are known as practical realizations of K-scheme, and some extrapolations of these methods are proposed. The new method is higher order than these methods without the use of extrapolations, and the author applies it to a financial problem of derivative pricing.

View full abstract

This paper studies a derivative on two assets when there is a model risk. We consider a two-dimensional Black Scholes model whose parameter processes are two drift processes, two volatility processes and one correlation process. It is assumed that true models of all parameter processes are unknown. Usually it is impossible to find a unique price of the derivative in this framework. Our problem is to find the price bounds of the derivative. We show a partial differential equation which the maximum price or the minimum price satisfies. To calculate the price bounds numerically, we propose a two-dimensional trinomial model and we study the model risk of options and their portfolio by a numerical analysis.

View full abstract

In this paper, we consider effects on hedging errors of a first-to-default swap (FTDS) under a hazard model with deterministic intensities from parameter estimation errors. Here parameter means default intensities of FTDS reference companies. As our mathematical model of FTDS and a credit default swap (CDS) market, we mainly follow Bielecki et al. [1] and [2] and use copula functions to represent the correlation of default times. We give some analytical and numerical results of hedging errors from parameter estimation errors.

View full abstract

This paper investigates the effect of “quantitative easing monetary policy (QEMP)” which the Bank of Japan (BOJ) adopted from March 2001 through March 2006, by changing operating target for money market from interest rate to the monetary base that is defined as the sum of “Cash” and “Reserve Deposit at the BOJ”. The paper confirms that the monetary policy has contributed to the recovery of the prolonged deflation. First we comparatively investigate economic activities in the usual economy period of (1981,1998) and in the zero interest rate period of (1999,2006), where vector autoregressive (VAR) model of (call rate, exchange rate, stock, nominal GDP, price) is estimated with “call rate” replaced by “Reserve” in the latter period. A monetary easing policy is effective through transmission path of stock market in both periods. Next we decompose money stock into transaction money and precautionary money to evaluate the transmission mechanism of the effect of reserves on the real economy by taking into account the financial anxiety. We have found a quantitative easing shock firstly increase “precautionary money” and secondly raise “Tankan”, which dispel the anxiety, and finally attain to “transaction money” and output.

View full abstract

The main theme of this research is numerical verification of applicability of a higher-order approximation to pricing barrier options, which is a both mathematically and practically important path-dependent type problem in mathematical finance. The authors successfully extend two types of algorithms called NV and NN. Both algorithms are based on a higher-order approximation scheme called Kusuoka approximation and have been shown to attain second-order approximation of stochastic differential equations as long as applied to European-type problems. In extending the algorithms, the authors apply a function representing the probability of hitting a boundary. In the numerical experiments, these two algorithms are compared with the Euler–Maruyama scheme which is one of the most popular first-order approximation schemes. As a result, it is demonstrated that the speed of calculation of these two algorithms could be much higher than that of the Euler–Maruyama scheme extended by the same way. It is concluded that one of the keys to improvement of the results is the construction of the function for calculation of the probability of hitting a boundary.

View full abstract

We propose a stochastic filtering model for Limit Order Book (LOB) where the signal process obeys the stochastic differential equation (SDE) whose trend function depend on the history of the LOB in such a way as to reflect the predicted direction of the mid price movement via the Support Vector Machine (SVM). Stochastic dynamics of the LOB is modeled with a queuing system where incoming orders and cancellations of existing orders arrive according to conditionally independent Poisson processes. Finally, numerical example with Nikkei 225 futures are presented.

View full abstract

In this paper, we consider effects of information, estimations and constraints on a portfolio optimization problem in mathematical finance. In particular, a portfolio optimization problem of investors who want to maximize their expected utility of their terminal wealth is considered. As our risky security model we adopt a factor model in which the growth rate depends on an exogenous factor. We assume several strategies whose differences come from information of markets, estimations of the factor and constraints of strategies, and we numerically investigate their effects on the expected utility. The logarithmic utility function as the utility function showing risk aversion investors is used in their paper.

View full abstract

In recent years, big data are utilized in many industries. In this study, in order to analyze duties of the nurses, we performed experiments to collect the duties activity data of the nurses for a long term. We set 38 nurses as subjects and asked them to carry out duties while attaching a wearable small sensor device, and collected the acceleration data, meeting information between nurses and the nurse duties information. In addition, we collected the location information of the nurses by using infrared information and communication equipment at the same time. From various data collected, we analyzed intensity and positional information of duties activity of the nurse, meeting information and the duties information between nurses and considered the influence that each factor affected to the nurse. As the result, we found that intensity of the activity increases in such nurses as who has many times of meeting with other nurses, visits the patient room many times, or who works on jobs concerning with the assistance of the patients such as rehabilitation assistance duties or the activity assistance duties.

View full abstract

Time decay of memory is measured using social media data. The behavior of decay of memory is known to be different for individual personal memory and for social memory. For personal memory, the decay is power law like. However, the decay of social memory seems to be exponential like decay, at least for popular movies. This result is significant to analyze social phenomena using sociophysics approach. Moreover, we found that the decay form is mixed decay of exponential and power law like for the social scandal topics.

View full abstract

Predicting short-term solar irradiance is important to introduce solar power in an electric power system. Considering that large quantities of solar power will be introduced across wide areas in the future, this paper addresses multi-point predictions of solar irradiance. Support vector machines (SVMs) perform well for predicting solar irradiance, but involve relatively high computational complexity. In addition, prediction models should be updated regularly to adapt to the season. Here, we introduce a multi-point prediction system that reduces the amount of calculation. Rather than constructing prediction models at all grid points, this system creates certain clusters based on the dynamic time warping (DTW) distance, and then constructs a small number of models representing each cluster. Mathematically, this consists of three algorithms, DTW, cluster analysis and SVM.

View full abstract

In this paper, we consider the Euler-Maruyama approximation for multi-dimensional stochastic differential equations (SDEs) with càdlàg process. We provide the strong rate of convergence when the drift coefficient is the sum of a Lipschitz continuous function and a monotone decreasing Hölder continuous function. We also prove the pathwise uniqueness and the existence for SDEs.

View full abstract

In this paper, we are concerned with a problem of optimization of the linear observations which are used in the stationary Kalman-Bucy filter. Especially, we consider the optimization of the gain matrix in the observation. In the previous works of the author, the condition of optimality was already shown by introducing a symmetric matrix called SSNRM which is a quadratic function of the gain matrix weighted by the inverse of the noise covariance matrix. In this paper, a new numerical algorithm to compute the solution is proposed which is applicable to the observations with lesser dimensions than the signal. Most part of numerical computations of this algorithm consists of that of the solutions of Riccati and Lyapunov equations which are easily implemented. The value of SSNRM is updated by a simple scheme based on the condition of optimality. The results of numerical experiments are provided to show the efficiency of the algorithm.

View full abstract

In this study, pattern dynamics on several coupled chaotic system with coupling strength by stochastically changing was investigated. In the previous studies concerned with this kind of work, almost of them are treated as a unique value and fixed for the coupling strength. We proposed a stochastic model of the coupled chaotic system when the coupling strength is changing based on a normal distribution, then various interesting phase patterns appeared in the proposed system. Several phase patterns, dynamics of the pattern generation and its statistic analysis of phase synchronization by computer simulation were shown.

View full abstract

As described in this paper, we derive the dynamics of closed-loop mechanical systems for underwater applications and present a numerical simulation of an underwater system. The dynamics model derived here is useful for analysis, control, and optimization for underwater parallel mechanisms and underwater object manipulation by multiple arms or fingers. As a first step, a two-dimensional underwater manipulator system with a stationary vehicle is modeled in this work. The dynamics model is coupled with hydrodynamic effects. Using the derived model, we conduct underwater object manipulation based on a force control method for use with underwater manipulators to prevent their contact points from slipping on an object surface. Moreover, we discuss the usefulness of the model and the dynamics of a free-floating vehicle-manipulator system with a closed-loop structure.

View full abstract

The hysteresis phenomenon exists in magnetostrictive actuators. When the hysteresis nonlinearity exists in a controlled system, the system usually exhibits inaccuracies or oscillations and even instability due to the undifferentiable and nonmemoryless character of the hysteresis. This paper proposes the robust control in broadband for magnetostrictive actuators, where the hysteresis is described by the Preisach model.

View full abstract

As described in this paper, we propose a robust tracking system with a stereo vision, IMU, and magnetic sensor for underwater robots to estimate the position of a stationary object even when the object is out of the field of view (FOV) of a camera or occlusion occurs. During maneuvering of underwater robots, they are subject to external disturbances. Their cameras are always moving. Therefore, objects in an image move out of the FOV of a camera unexpectedly. As described in this paper, we integrate an IMU and a magnetic sensor with a stereo vision sensor to estimate an object's position robustly. In the process of the estimation, Kalman filters are used to estimate the sensor system orientation. First, we conduct numerical analyses to investigate the estimation accuracy of the proposed method against sensor noise. Then, a preliminary experiment is conducted to demonstrate the robustness of the proposed method against occlusion and moving out of the FOV. As a first step, we examine the tracking performance with regard to rotational motion of the sensor system.

View full abstract

In use of a multi-link flexible manipulator in general environments such as homes and offices, collisions between the flexible manipulator and various obstacles and/or working humans is unpredictable and brings serious damage to the manipulator and humans or obstacles. For example, when the flexible manipulator is used for assistant of the meal, humans or obstacles can be moving in the work space of the flexible manipulator. It causes the collision between the flexible manipulator and the unknown obstacles and there are the cases which cannot be avoided the collision. In order to reduce the damage of the manipulator and the humans and then to re-plan the path plan to the desired position, the unexpected contacts or collisions should quickly be detected and the collided link should be identified for reacting to suspend the motion of the manipulator and for planning new path. This paper presents a method of collision detection for each link of a two-link flexible manipulator by introducing respective collision detection function based on residual processes generated by the innovation process of a nonlinear state estimator. Furthermore, the mathematical structure of collision detection for each link is investigated.

View full abstract

In this paper, we show the results on the ethylene gas detection. The ethylene gas is generated as one of the by-products of the breathing in many plants. It may become harmful to the freshness when we keep on having collected around a ripe product excessively. Ethylene gas accelerates a process to mature and caused self-acceleration of the deterioration. The main promotion of ethylene is extremely a small amount, but gas chromatography and mass spectrometry are used as the measurement devices which can separate gases. The components are expensive and need much experience for operation. We use a neural network based on the features that can be obtained from measurement data assuming the real environment such as steam, alcohol, and ethylene. Using a small odor measurement system that created using the mass sensor of the inexpensive crystal oscillator can be detected selectively to the ethylene gas. We consider the technique to implemente and configure the system to show the processing result.

View full abstract

This paper considers an estimation of rice-planted area by using remote sensing data. The classification method is based on a competitive neural network and the remote sensing data are observed by a satellite before and after planting rice in 1999 in Hiroshima, Japan. Three RADAR Satellite (RADARSAT) and one Satellite Pour l’'Observation de la Terre(SPOT)/High Resolution Visible (HRV) data are used to extract rice-planted area. Synthetic Aperture Radar (SAR) backscattering intensity in rice-planted area decreases from April to May and increases from May to June. Thus, three RADARSAT images from April to June are used in this study. The Self-Organizing feature Map (SOM) classification was applied the RADARSAT and SPOT to evaluate the rice-planted area estimation. It is shown that theSOM of competitive neural networks is useful for the classification of the satellite data.

View full abstract

In this study, the motions involved in hula hooping were divided into several chronological segments, and a method for determining the skills involved in each segment developed. A significant number of the actions we perform every day, such as walking and running, involves periodic operations. These actions are learned early in childhood and, in later life, we go through the motions unaware of the component actions. However, when it becomes necessary to reacquire these talents through rehabilitation or other means, a considerable amount of effort is subsequently required through repeated trial and error before we can efficiently perform them once more. We have developed an algorithm that can chronologically and quantitatively represent the skills involved in these periodic behaviours. This paper outlines the construction of a Bayesian networks model used for analysis, and presents information on the number of segments, representation of the skills involved in a motion/action when there is variance in the position of a segment, and the reliability of those skills.

View full abstract

In recent years, informed consent has become more and more important in every area of medical care. It is expected that new techniques for recognizing MR images will be developed. The Medical doctors look for indications of disease by using a color system in MR images and then they explain it to there patients. However, it is often difficult for patients to understand the explanations. In this case time, the use of AR (augmented reality) will be appropriate. Introducing this technique to patients will help then to recognize and understand two pictures of the MRI system. The method is based on the ideas proposed by our team of researchers. The MR images give two kinds of weighted images: T1 and T2. We have developed a make the system which extracts tissue by comparing both weighted images. We can extract a specific tissue by setting the threshold value. The extracted tissues are mapped and superimposed over the original images. The AR contents are extracted from samples of water and fat. We are currently using software called the software “Metaio Creator”. The image of the brain is the AR marker. The software recognizes the image and shows an image which is easier for the patient to understand. We have succeeded in creating a visualization of AR. We can see an image of the extracted sample by looking at the tablet. The brain image markers are displayed.

View full abstract

In this paper, we consider the algorithms for minimizing the total variation norm of an image for zooming a single image. Super-resolution imaging is a technique to reconstruct a high-resolution image from a low-resolution image. To obtain a high-resolution image that preserves edges and has no ringing, we choose the image which minimizes the total variation (TV) norm of the image with constraints, according to Chambolle's method. This regularization based method is compared with some traditional methods. On the other hand, a total variation based method has huge computational cost because of it's iterative scheme. We also compare processing times in this paper.

View full abstract

In this paper, practical methods for improving GPS relative positioning accuracy in long baseline are presented and examined by using real data. In our previous work, the method for improving positioning accuracy has been presented by applying PAR(Partial Ambiguity Resolution) method. However, it is also shown that the estimate of the position as well as the integer ambiguity are unreliable due to ionospheric delay, satellite orbit error, etc. Therefore, in this paper, we propose a new long baseline positioning method by applying PAR method and LEX(L-band EXperiment) correction information. LEX correction information is a signal broadcast from QZS(Quasi-Zenith Satellite) which is Japanese navigation satellite. By using PAR and LEX correction, the experimental results show that the dominant error sources such as the ionospheric delay and satellite orbit error are corrected by LEX signal, and the integer ambiguity can be more efficiently resolved. Consequently the positioning accuracy are improved.

View full abstract

Quasi-Zenith Satellite System (QZSS) is a system that improves performance of Global Positioning System (GPS) in urban areas of Japan. QZSS is designed to transmit the augmentation signals so-called the LEX (L-band Experiment) signal. In this paper, we propose a method to apply the LEX signal and Least-squares AMBiguity Decorrelation Adjustment (LAMBDA) to Precise Point Positioning (PPP), and the positioning performance as well as the accuracy of correction information of LEX the signal are evaluated.

View full abstract

In this paper, we present the regional ionospheric delay models for Global Navigation Satellite System (GNSS). An ionospheric delay is a dominant error on GNSS positioning. Then, an ionospheric delay should be corrected for a high positioning accuracy. In this study, the methods to develop the ionospheric delay models are proposed. Ionospheric delays can be measured by the geometry-free combination of the GNSS regression (GR) models of dual frequency bands. An ionospheric delay is proportional to total electron content (TEC) and then a vertical component of TEC (VTEC) is modeled. The VTEC models have been developed based on spherical cap harmonic analysis (SCHA) by using GNSS earth observation network (GEONET) data. The GEONET data are observed only in Japanese region and provided by the Geospatial Information Authority of Japan (GSI). We also investigate the prediction of ionospheric delays with the post-processing models developed by the SCHA. We propose the prediction methods by periodic autoregressive (PAR) models.

View full abstract

Existing smell classification devices can display the strength of smells as quantitative values by using electric circuits. But they cannot perform the type of smells such that they are good or bad. In this paper, we construct a smell measurement system that can classify the type of smell by using neural networks and metal oxide semiconductor gas sensor arrays. In particular, we classify the kind of wines, which has been considered as the most difficult classification. We measure the smell data of white wine and red wine by using a method to change temperatures of smells of wine data and classify them by using a neural network. Using the classification rate of white wine and red wine, we consider the ability of classification of the proposed method.

View full abstract

The empirical evaluating formulae of the stochastic differential equation (SDE) are considered in this study, in order to reconstruct the information of the dynamics by the use of time-series itself. In this framework, we regard the time-series as a realisation of a solution of the SDE. By using the solution-coeffcient relation studied in the probability theory, we can evaluate the drift vector and the diffusion matrix of the SDE, which are calculated by the first and the second order variations per unit time. The drift vector represents the deterministic bahaviour of the motion and the diffusion matrix play the role of stochastic effect. Utilising the statistically evaluated SDE, one can obtain the predictability estimates in terms of the ensemble variance. The results indicate that the deterministic components are significant for the short-term prediction. Moreover, we need the stochastic components to obtain better predictability estimates by the use of the numerical simulation of the SDE.

View full abstract

The smartphone market has been growing rapidly and various smartphone content services are provided. However, the lack of bandwidth for the mobile network limits the smartphone video services to real-time. Most of the videos for smartphones on the internet are low-bitrate and blurring. In our previous works, we proposed a new SR technology that uses nonlinear signal processing (NLSP) and we demonstrated its theoretical capability and image quality for 4K TV sets. We applied NLSP to a smartphone. However, the capability of the NLSP on the smartphone display has not been tested. A subjective assessment using best-worst method was performed to verify the performance of NLSP on the smartphone display. The assessment was conducted by comparing the image qualities of the smartphones with and without NLSP and different manufacturers’ smartphones. Assessment data were statistically analyzed, and the assessment results proved that the image quality of the smartphone with NLSP is most superior in resolution.

View full abstract

The aim of this research is to evaluate the behavior of small scaled wind turbine system against strong wind input using electromagnetic stall control system. In general, the wind turbine system is generating energy from the revolution of blades. The revolution of blade is varying through the environmental wind that is swept to blade. Therefore, if there is strong wind then the revolution of blade is increased and consequently more energy will be generated. Essentially the wind turbine systems are located in the gale area in order to generate energy efficiently. However, there is a boundary of revolution of blade, especially in small-scaled turbine due to small inertia momentum. If the angular velocity exceeds the boundary of revolution then wind turbine system may breakdown. Thus in this paper, in order to avoid the malfunction of small-scaled wind turbine system, electromagnetic stall control (ESC) is suggested. ESC can control the angular velocity without having any connection with shaft. Thus, we would control angular velocity much efficiently than the conventional stall control method such as friction stall control. For stability determination and performance evaluation phase plane method is applied in order for certain determination of stability. As a consequence, we could verify the reliability of ESC system by phase plane method.

View full abstract