A necessary and sufficient condition such that there exists a proper stable rational transfer function whose H_∞ norm is less than or equal to a specified value is derived as a matrix inequality composed of its input-output covariances. Then, it is applied to a model invalidation problem with additive bounded uncertainty. A numerical example demonstrates that the model invalidation condition based on input-output covariances is robust against additive measurement noise.

View full abstract

Blind signal separation based on statistic independence has been studied extensively. Particularly difficult is seperation of convolutive mixture. This can be solved approximately by a method with an FIR filter which is a cascade connection of many delay elements. It is, however, more desirable if the signal can be deconvoluted with an IIR filter (having feedback loop) of a modest order. This paper proposes such a demixer by projecting the conventional learning law based on the FIR approximation toward the parameter space of IIR filters. Numerical simulation is carried out to evaluate the proposed method.

View full abstract

For computing a Nash (saddle point) solution to a zero-sum differential game for a general nonlinear system, Mukai et al. presented an iterative Sequential Quadratic-Quadratic Method (SQQM) as follows. Given a solution estimate, they defined a subproblem which approximates the original problem up to the second order around the solution estimate. They proposed to replace the subproblem with another subproblem in order to obtain a game problem with only a linear dynamics by removing the quadratic terms in the system dynamics and adding them to the payoff function as in Lagrangian function. We can now solve this subproblem conveniently by a Riccati equation method. We then update the solution estimate by adding its Nash solution to the current solution estimate for the original game. Through our extensive experiments, we observed not only local convergence of the SQQM but also much faster convergence of the SQQM than the iterative methods based on lower order approximations such as the Sequential Linear-Quadratic Method (SLQM). In this paper we will establish local convergence of the SQQM.

View full abstract

This paper describes a method of collision detection and control of a single-link flexible arm using a couple of state estimators. In order to improve the control performance of the flexible arm with the innovation based collision detection system when unlooked-for obstacles collide with the side of the arm, we introduce a robust optimal filter, which decouples undesirable effect of the collision force from the estimation error. The controller generates a control torque based on the state estimate using the robust optimal filter. Collision detection is then realized by using the innovation of a Kalman filter. The proposed system enables to avoid that the control performance is deteriorated when collisions occur.

View full abstract

In this paper, we consider an optimization problem for observations of stationary LQG stochastic control systems which employ the stationary Kalman filter. The performance of the Kalman filter and that of the LQG stochastic optimal control are both dependent on the gain matrix in the linear observation. We have already developed methods of optimizing this gain matrix based on the estimation or control individual performance under a quadratic performance criterion. This paper discusses a hybrid problem by taking into account of both estimator and regulator performances. By introducing the eigenvalues-eigenvectors representation of a nonnegative definite symmetric matrix, the condition of optimality is derived. Also, numerical calculations are easily carried out by introducing multi-dimensional polar coordinates systems.

View full abstract

For a class of second-order mechanical or structural systems subjected to random disturbances, their discrete-time models are derived. Since the output of stochastic system is random and has no longer its velocity in the ordinary sense, the notion of the Nelson process which is a stochastic process having forward and backward mean velocities is introduced to describe the action integral. Then the stochastic version of the Euler-Lagrange equation is derived from the Hamilton's principle. The discrete-time second-order model is obtained by discretizing the stochastic Euler-Lagrange equation and/or Hamilton's principle. The structural matrices are identified using the discrete second-order model.

View full abstract

Motivated by quantitative evaluation of aging properties of electrical cables, a computational method for identifying physical parameters of dielectric materials is considered. First, a nondestructive testing using micro-waveguide is described by electric and magnetic fields defined on the problem. Secondly, an inversion technique based on Markov Chain Monte Carlo method (MCMC) is employed for solving the electromagnetic interrogation treated here. Finally, a numerical scheme of the direct problem is developed using the finite-difference time-domain method (FDTD) in two spatial dimensions. Applying Metropolis-Hasting algorithm to the inversion presented here, results on computational experiments are demonstrated in order to show the feasibility and applicability of the proposed method.

View full abstract

In this paper, a method of simultaneous identification of the unknown parameters involved in the system matrix and unknown magnitude of suddenly discharged exogenous input is proposed for a class of linear stochastic systems using the idea of pseudomeasurement which has been developed extensively by the authors. The principal attack is to augment the pseudomeasurements with the original observation process and construct the Kalman filter. Numerical examples are presented to show the efficacy of the proposed approach.

View full abstract

Infectious diseases remain a threat to mankind in the 21st century. A prevention of prevalence of infectious diseases is one of important problems in epidemiology. In the past, the public health system has prepared for some strategies such as antibiotics and vaccines to control the infectious disease development. In order to build up more effective strategies, we need a precise theoretical analysis of infectious diseases. From a theoretical viewpoint, the mathematical model which describes the spread of the infectious disease has a very important role. In this paper, we study the stochastic modeling of the infectious disease in populations consisting of four populations: susceptible, infected, recovered and vaccinated ones. We consider the optimal vaccination strategy under the framework of the stochastic optimal control problem. Applying the four-step scheme to the considered stochastic optimal control problem, we construct a feasible optimal vaccination system.

View full abstract

In this paper we study hybrid estimation for linear continuous-time systems with noises not to be restricted to be Gaussian. It is assumed that modes of the systems are not directly accessible. We consider optimal estimation problems to find both estimated states of the systems and a candidate of the distributions of the modes over the finite time interval. We adopt most probable trajectory (MPT) approach. Q. Zhang (2000) has presented hybrid filtering algorithm, i.e., causal estimation, by MPT approach. We consider both filtering and smoothing problems in this paper. We also consider the cases that mode transitions follow Markovian jump processes and present nearly optimal estimation algorithms for limiting mode distributions. We can expect better estimation performance by taking into consideration noncausal information of observations. The hybrid smoother is realized by two filters approach.

View full abstract

In order to design tracking control systems for a class of systems with rapid or abrupt changes, it is effective in improving tracking performance to construct preview control systems considering future information of reference signals ([12]). In this paper we study the H_∞ tracking problems with preview by state feedback for linear continuous-time Markovian jump systems on the finite and infinite time intervals. In the case of infinite time horizon the necessary and sufficient conditions for the solvability of the H_∞ tracking problem with preview are given by coupled algebraic Riccati equations (CAREs) or coupled linear matrix inequalities (LMIs).

View full abstract

In this study, a revised Group Method of Data Handling (GMDH)-type neural network using principal component-regression analysis is proposed and applied to the nonlinear system identification. GMDH-type neural networks can automatically organize neural network architecture by heuristic self-organization method and structural parameters such as the number of layers, the number of neurons in hidden layers and useful input variables are automatically selected so as to minimize the prediction error criterion defined as Akaike's information criterion (AIC) or Prediction sum of squared (PSS). But, in the heuristic self-organization method, the multicolinearity generates and the network architecture becomes unstable. In this study, the principal component-regression analysis is used for estimating the parameters of the neurons and stable and accurate multi-layered architectures of the GMDH-type neural networks are organized using the heuristic self-organization.

View full abstract

A revised Group Method of Data Handling (GMDH)-type neural network algorithm for medical image diagnosis is proposed and is applied to medical image diagnosis of liver cancer that is called hepatocellular carcinoma (HCC). In this algorithm, the knowledge base for medical image diagnosis are used for organizing the neural network architecture for medical image diagnosis and the revised GMDH-type neural network algorithm can identify the characteristics of the medical images accurately. The optimum neural network architecture fitting the complexity of the medical images is automatically organized so as to minimize the prediction error criterion defined as Prediction Sum of Squares (PSS) and it is shown that the revised GMDH-type neural network can be easily applied to the medical image diagnosis.

View full abstract

In this paper, a direct adaptive fuzzy control using supervisory compensator is designed for a MEMS gyroscope. The parameters of the membership functions change according to the adaptive law for the purpose of controlling the system to track a reference trajectory. A fuzzy controller that can approximate the unknown nonlinear function and compensate the system nonlinearities is incorporated into the adaptive control scheme in the Lyapunov framework. A supervisory compensator is adopted to guarantee the stability of the closed loop system. Numerical simulations for a MEMS angular velocity sensor is investigated to verify the effectiveness of the proposed adaptive fuzzy control scheme and demonstrate the satisfactory tracking performance and robustness.

View full abstract

In this paper, the author investigates a new class of fuzzy random sets as the vague perception of a crisp phenomenon or a crisp random phenomenon. First, considering that the vague perception of a crisp phenomenon fluctuates slightly but randomly by the state of a capricious person's mind, a new class of fuzzy random set is introduced. Secondly, the fuzzy random set is generalized as the vague perception of a crisp random phenomenon. The expectations of introduced fuzzy random sets are investigated from the viewpoint of the multivalued logic proposed by Kwakernaak[1].

View full abstract

This paper presents a load balancing routing protocol β-OMP with Bayesian learning scheme. Load balancing protocols are important to optimize property for traffic engineering and they can utilize Internet resources effectively. However not so many of load balancing routing protocols are available to the Internet. Optimized Multipath(OMP) is one of load balancing routing protocols, which is considered to be most applicable for the Internet, although it has slow convergence property and may cause congestion of particular links due to its probabilistic traffic regulation manner. Hence we propose the β-OMP, which is introduced Bayesian learning scheme to the OMP and show its efficiency through several simulation results under some traffic models.

View full abstract

This paper considers the problem of designing optimal smoothing vector spline curves with equality and/or inequality constraints cross-coupled among the element curves. The splines are constituted employing normalized uniform B-splines as the basis functions. Then various types of constraints are formulated as linear function of the so-called control points, and the problem is reduced to convex quadratic programming problem. The performance is examined by some numerical examples.

View full abstract

In this paper, we present a method of detecting abnormal measurement data due to cycle slips and multipath in GNSS (global navigation satellite system) for RTK-PPP (real time kinematic - precise point positioning), based on GR (GNSS Regressive) models. After detecting the received signals with cycle slips and multipath from the GNSS satellites, we exclude psuedoranges data for multipath and re-estimate the integer ambiguity included in L₁ or L₂ carrier phase for cycle slips, for maintaining the accuracy of GNSS positioning. We present two methods of detecting cycle slips. The first method is based on statistical testing of the innovation process and the second method is using difference between successive-time estimates of integer ambiguities, with applying the Kalman filter. Finally we present the experimental results using real GPS observables in the automobile and show the feasibility of detecting cycle slips and multipath by our method proposed in this paper.

View full abstract

In this paper, the combined GPS and GLONASS Precise Point Positioning (PPP) method based on GR (GNSS Regression) model is described. The performance of the combined GPS and GLONASS PPP is assessed using observation provided by the receiver. The corresponding positioning accuracy and convergence time are compared for GPS-only and GPS/GLONASS precise point positioning. The results have indicated a significant improvement on the position convergence time although the improvement is quite dependent on the geometry improvement of the visible satellite configurations. The results also indicate a slight improvement on the positioning accuracy due to limited availability of GLONASS satellites.

View full abstract

We compare several approaches to uncertainty propagation that have been used in the literature to formulate the uncertainty in a dynamical system governed by ordinary differential equations. Specifically we focus on the evolution of probability density functions of the associated stochastic processes, and discuss their applications in different fields.

View full abstract

The authors have been developing a mobile robot with sensors for various services in the university campus. A prominent feature of university campus is a substantial amount of pedestrians in the outdoor environment. This feature is also typical in the shopping streets where cars are shutout. Motivated by such facts, we propose a detection procedure of static and mobile objects in the environment, where probabilistic model is employed. In this paper, we implemented the detection algorithm in the offline mode. We demonstrate the detection experimental results, which validate the usefulness of the proposed algorithm.

View full abstract

In this study, a framework of fusion network with several sensors for an inverted pendulum mobile robot is investigated by developing various services as a campus robot in our university. Several devices has been equipped on the robot for sensors as detecting obstacles, walls, human body, boundary of the ground and so on. In order to implement an algorithm for safety to drive, a framework of sensor fusion network realized by RT-middleware components is explained. As one of example, we show an experimental result using the framework on the campus event.

View full abstract

Mobile robot localization problem is determining a robot's pose in a given environment map. This problem is separated into three categories (position tracking problem, global localization problem, kidnapped problem) based on complexity. This paper presents a new evolutionary localization algorithm for global localization problem. Global localization problem aims to determine the robot's pose in a known environment without initial robot's pose information. In this study, we use differential evolution (DE) as evolutionary computation and define the global localization problem as a global optimization problem. Several studies have reported that differential evolution has an excellent ability for solving global optimization problems. The proposed method has been tested in some simulated environments to demonstrate the effectiveness.

View full abstract

This paper describes a method for localization of a mobile wireless LAN client in multistory buildings using the strength of received signals in a state space framework. Data pertaining to the physical positions of personal electronic devices or mobile robots are important for information services and robotic applications. We focus on integration of the estimation results with other sensor data based on a state space framework. The estimation model for location provides the variance of the position of the mobile client. The estimation model is re-initialized when the mobile client moves to the other floor in the building based on the detection of the change of number of the floor where the mobile client moves by the Bayesian inference. Experimental results show the feasibility of the method.

View full abstract

This paper describes a method for sensor data fusion of thermal imaging sensor and a vision system such as a camera for probabilistic target tracking for autonomous robots. Person tracking is a crucial issue for achieving tasks for communication with people and following the target person by the robots. The integration of multiple kinds of sensor data for validation of the recognition result is desirable because of the uncertainty of recognition results based on the vision system. We use a stochastic model based on asynchronous updating the model of the tracking target by measurement data of the camera and the thermal imaging sensor. Experimental results indicate the feasibility of the proposed method.

View full abstract

Some numerical results with respect to sensitivity analysis are given under a linear stochastic differential equation with long memory by using fractional Brownian motion. They are related to Greeks calculations in mathematical finance. In order to simulate them, formulas driven by the Malliavin calculus are used. Those results are the same as in the case of standard Brownian motion. That is, in the case of discontinuous payoff functions, results of the Malliavin method are stabler than ones of the finite difference method. Also they are true even in 2-dimensional situations for discontinuous payoff functions.

View full abstract

Y.Hishida and K.Yasutomi[4] give a numerical method for pricing path dependent options. This method uses the Markov property and the temporal homogeneity property for each underlying asset process. This enables us to approximate a path dependent option by using some kinds of plain vanillas. This approximate method is valid for diverse processes and path dependent options. Y.Hishida and K.Yasutomi[2] give a numerical examination for Asian option when the underlying asset behaves as geometric Brownian motion. In this paper, we give a numerical examination for Asian option when underlying asset behaves as geometric normal inverse Gaussian process.

View full abstract

In this paper, we propose a resotration algorithm of rotational motion blurred images using inverse filters. In paticular, to overcome this problem, we focus attention on the two methods. First, we interpolate the pixels near the rotational paths by shift-invariant operation. Second, we reconstruct the input image on the another domain.

View full abstract

A parallel distributed schemes is presented for generating as-is graph structure spanning complex local geographics. In this scheme, GPS tracks are considered to be uploaded from a multitude of probe vehicles to expand a network of feasible roadway patterns. By associating the GPS tracks with multi-Gaussian distribution conditioned by landmark distribution, the graph structure is successively augmented in terms of feasible roadway segments. On the augmented information, symbolic design of over-the-horizon maneuvering process can be interactively substantiated into a chain of topological minors grounded on the local geographics.

View full abstract

In this research, a teleoperation system is investigated by the viewpoint of lighting conditions. Communication channels are established by using the Internet technology. The teleoperation system consists of two different techniques. One is the force-free control (FFC), which is used to realize the passive motion of the robot arm for an external force. The other is an image processing component in which template matching (TM) techniques are adopted to realize the accurate motion by means of visual servo control. Since the template matching technology plays a vital role in our teleoperation system, it is important to investigate the behavior according to varying lighting levels. The experiments were performed to investigate the influence of lighting conditions by using actual robot arms with vision system.

View full abstract

BRDF measurement generally require delicate instruments and complex measuring processes for obtaining geometrical information of viewer, light source and surface. In this paper, a novel approach is developed to evaluate reflectance properties such as specular or diffuse albedos by applying Independent Component Analysis (ICA) to the images obtained by changing the orientation of the light source. Surface roughness is also evaluated as a relative value by using separated specular images. The proposed approach does not require any geometrical information and is very suitable for the materials whose diffuse and specular components have obvious differences, such as metal. The evaluated reflectance properties can be applied to rendering, diffuse-specular separation, 3D measurement of metallic surface or design of various kinds of image processing algorithms.

View full abstract

In this paper, in order to improve the real-time speech visualization system called “KanNon”, we work out experiment of phoneme intervals and vowel recognition. We show the structure of the system, then the feature of the speech signals is obtained by the AR model. Phoneme intervals and vowel recognition are obtained by the AR parameter estimation method.

View full abstract

In this paper, we propose a prefiltering method of frequency domain data for linear multiple-input multiple-output time-invariant system identification. This paper only deals with the case where the frequency domain input-output data is given by discrete Fourier transform. The concept of the proposed method is to eliminate the disturbance factor from the output data, it can be performed by orthogonal decomposition based on Wold's decomposition. Thus, by employing MIMO system identification method based on the frequency domain data given by the discrete Fourier transform, the data implemented by the proposed method can derive more accurate model. LQ decomposition is the key tool in the proposed method, it actually computes the orthogonal decomposition of the frequency domain data instead of Wold's decomposition. The efficacy of the proposed method is shown in numerical experimental.

View full abstract

We have recently proposed the adaptive velocity and acceleration estimators based on the adaptive control theory. In this report, we perform the experimental validation of proposed estimators with the Crossbow NAV440 that is a product of MEMSIC,Inc. Using Crossbow NAV440, we can obtain the roll, pitch, and yaw angles and their velocities and accelerations. The performances of the proposed differential filters are verified by using the measurement data with NAV440. Moreover, the adaptive velocity and acceleration estimators are compared with the approximate differentiatoｒ and the exact differentiator.

View full abstract

We study the identification problem for Heston stochastic volatility model, which is widely used as a model of a stock in finance. Based on observation data (log price), we estimate the stochastically moving volatility with its systems parameters by using particle filtering technique.In order to apply the particle filter algorithm, we need to convert the original continuous model to a discrete one. In this paper, noting that the volatility process of Heston model can be generated from the non-central chi-square distribution, the exact particle filter algorithm is applicable. We also demonstrated some simulation studies for showing the efficiency of our proposed algorithm.

View full abstract

In this paper, we perform an estimation of the volatility of the forward rate using the Fourier method proposed by Malliavin and Mancino [4], together with the classical Principal Component Analysis. We have observed similar anomaly with the one in Liu [2]. We also give a remark on the practical use of the Malliavin-Mancino method.

View full abstract

This paper provides some results of numerical experiments of the Put-Call Symmetry method proposed by the authors in [4] under Heston and λ-SABR models.

View full abstract

The Lévy process is considered to predict reliability assessments for degradation phenomena. In particular, the geometric Brownian motion and the gamma process models are handled, and the Bayesian procedures are applied to the models. The Markov chain Monte Carlo method and the empirical Bayesian method based on hierarchical Bayesian modeling are used to estimate the parameters. In addition, the maximum likelihood method and the generalized moment method are also appropriated to the models. The information criterion EIC is used to select a fitted model among the stochastic models and the estimation methods. The marginal likelihood with the Laplace's method plays important roles in the methodologies proposed in this paper.

View full abstract

In this paper, the multi-layer ionospheric model based regional VTEC (Vartical Total Electron Content) estimation methods applied by GNSS (Global Navigation Satellite System) Regressive models (abbreviated as GR models) are presented. In the methods, the SCHA (Spherical Cap Harmonic Analysis) as well as the SHF (spherical harmonic functions) are also applied to model the regional VTEC distributions. The ionospheric delays (or advances) over Japan are estimated by using GEONET (Gps Earth Observation NETwork) data provided from the GSI (Geospatial Information Authority of Japan). By applying the ionospheric correction based on the multi-layer model, the positioning accuracy can be improved, and consequently the regional ionosphere can be modeled more accurately by the multi-layer model.

View full abstract

Cycle slips in Global Navigation Satellite System (GNSS) signal tracking are significant error sources for carrier phase based GNSS applications, therefore it is necessary to detect and correct them. In this paper, we propose a method of cycle slip detection and correction with the time-differenced carrier phase measurements for single frequency applications. The time differential technique is applied to cancel the integer ambiguities, but any cycle slips remain in the time differenced carrier phase measurements. The two kinds of statistical hypotheses tests using residuals in the least squares adjustment are adopted to detect cycle slips. We have evaluated the performance of our proposed method with real data for different time intervals. Finally, we show the results of cycle slip detection and correction tests by using our proposed method with real receiver data with simulated cycle slips in static and kinematic cases. Our performance evaluation results show that our proposed method can detect and correct the cycle slips efficiently in static and kinematic cases within 20-second data gaps. These results are a promising outcome for single frequency applications which still lacks the accuracy and reliability compared to dual frequency applications.

View full abstract

The purpose of this paper is to analyze stability of the so-called particle swarm optimization (PSO) algorithm by modeling it as a Markov jump linear system with multiplicative noise, so that we can take into account the interaction between particles. A stability condition is represented in terms of linear matrix inequalities which is checked efficiently by using a convex optimization solver. A condition on the decay rate of the PSO algorithm is also shown. Stability region and contours of the decay rate of the PSO algorithm are shown by using the proposed analysis methods. Consequently, we provide more precise understandings of the PSO algorithm when the interaction between particles is considered.

View full abstract

There are many types of Evolutionary Algorithm. Among them, Genetic Programming (GP) has the rich expression of solution for a lot of problems, and Genetic Algorithm (GA) is good for the tuning of parameters on array. However, when we apply each of them to the extraction of investment rule, it needs a lot of searching for useful solution in the field of foreign exchange. So we introduce the diversified investment using multiple solutions, and even if the searching does not work well, we can decrease the risk with the diversified investment. In addition, to increase the success rate of searching and applying, we discuss about the hybrid algorithm of GP and GA, and we show its effectiveness in the extraction of investment rule. In numerical example, it is shown that our proposed model have better performance rathar than the usual models combining GA and GP.

View full abstract

In this paper we develop an optimization problem for assignment of jurisdiction area to target facility. We adopt the minimization of welfare loss of customer as the criteria of evaluation. We refine usual formulation of optimization problem for facility location of public sector by introducing the concept of geographical advantage. Such idea leads originality of this study. In numerical example, we show our proposed method can support decision making of assignment problem in the real world by using program with the Google map API.

View full abstract

Swarm Intelligence is the optimization technique based on the action of groups, such as a bird and a fish, and a colony of the ant. PSO which is one of the technique is developed and applied in various studies. However, convergence of PSO is groundless. From the point of view, in this paper, we have proposed a new hybrid dynamical system of Swarm Intelligence and Neural Network dynamics for finding better optimal solution. As the main results of this paper, we first show how to combine PSO and neural network mechanism by the theoretical analysis, and confirmed that proposed system can realize interpolation search based on global information of objective environment. The meaning of the interpolation search is the mechanism of the dynamics of local search which is included in global search of PSO.

View full abstract

During 2001 through 2006, the Bank of Japan (BOJ) adopted ”quantitative monetary easing”. Although zero interest rates seem to cause the ”liquidity trap” in (2001,2006), the growth rate of GDP can be seen to increase gradually. We shall show empirical studies such that there is a transmission path where the increased ”Reserve at the BOJ” makes the money turned to the stock market, and hence GDP is increased out of the ”liquidity trap”. The principal line of attack is to decompose money into precautionary money demand and transaction one. Assuming that precautionary money demand is a function of GDP and fluctuations due to business cycle, we can obtain a new estimation result of precautionary money demand. Defining ”adjusted money＝m2cd - precautionary demand”, we can show a transmission path such that ”Reserve at the BOJ” → adjusted money → stock prices → GDP. Six variables of (adjusted money, stock, exchange rate, interest rate, price, GDP) are combined in a macro money system.

View full abstract

We can utter sounds by changing a shape of our vocal tract. A vocal tract characteristic has been analyzed by AR model for speech signals and a vocal tract area function can be estimated by PARCOR coefficients computed from the AR model. On the other hand, a method has been developed for three-dimensional reconstruction of the vocal tract shape from magnetic resonance imaging (MRI). In this paper, the vocal tract area function during vowel production will be estimated based on the following two methods and they are compared.

　(1) It is estimated as PARCOR coefficients from AR model for speech signal.

　(2) It is calculated as cross section area of vocal tract extracted from MRI data acquired while a subject is uttering the long vowel.

View full abstract

The problem of robust stability for a class of uncertain Markovian jump stochastic systems with mode-dependent time-varying delays is studied. By constructing Lyapunov functional and applying Itô's differential formula, novel delay-dependent stability sufficient conditions are derived based on linear matrix inequality (LMI). Two numerical examples are presented to show that the correctness and less conservativeness of the proposed method.

View full abstract

In this paper, chaotic behavior in a piecewise continuous map(PCM) associated with sine circle maps of nonlinear systems of Van der Pol-Mathieu type is studied. At first, a sine circle map and its approximate PCM are derived. Secondly, based on computational experiments on Lyapunov exponents to approximate PCM, existence conditions are shown to generate chaotic oscillation and periodic oscillation in PCM. Finally, it is shown that the rotation numbers of oscillation in PCM form a part of successive Farey series and form devil's stairs in the region of cut off parameter such that Lyapunov exponent is positive. Furthermore, it is shown that the order of appearance of the oscillation in PCM has a self similar hierarchy.

View full abstract

Diffenential flatness theory and Extended Kalman Filtering (EKF) are used for implementing a sensorless control scheme for induction motor drives. It is shown that the induction motor is a differentially flat system, since all state variables of the circuits describing the motor's dynamics can be expressed as functions of the flat output and its derivatives. The flat output consists of the angle of the rotor and of the the angle of the magnetic field. Flatness-based control for the complete (sixth-order) induction motor model is analyzed while the Extended Kalman Filter is proposed to estimate the state vector of the nonlinear electric motor using a limited number of sensors. The efficiency of the above mentioned Extended Kalman Filter-based control scheme, is tested through simulation experiments.

View full abstract

An algorithm for music transcription system that displays notes and rests on the five lines of music sound is proposed. It is able to process multi-voice music which includes chords, superimposed melodies etc. The system first projects the sound onto the time-frequency plane applying Gabor wavelet transform, and then identifies tempo, the length of notes etc. with referring autocorrelation of the projection. At next, it estimates actually played pitch names that correspond to the fundamental frequencies of simultaneously played tones. It applies a state estimation technique to cope with considerable harmonics which depend upon instruments.

View full abstract