We propose a 3D object recognition method using inverse synthetic aperture radar images, which doesn’t need any aspect information. It employs a simple support vector machine of linear kernel. Numerical experiments clarified that when the movements of the targets are ideally / practically compensated, it recognizes cones and cylinders with as much as 95% accuracy.

View full abstract

For analyzing a sample textile fabric structure, a novel method is proposed to obtain positional information of each yarn of the sample from voxel data made out of its X-ray computer tomography (CT) images. The positional information is obtained by tracing the each yarn focusing on the fiber orientation. The effectiveness of the proposed method is confirmed experimentally.

View full abstract

We describe a new laser rangefinder that accurately determines the 3-D shapes of specular objects with certain types of background light through the use of a neural network. Our rangefinder has following features: (1) the capability of detecting incident position and angle simultaneously, (2) the robustness against light noise present in the environment in which the measurements take place.

View full abstract

This paper describes an extended 3-D vision system for spherical objects using highlight patterns formed by specular reflection on surfaces of the objects. The extended system has less restriction about arrangement of cameras and point light sources than the former system that proposed by the authors. The basic performances of the extended system are proved by experiments.

View full abstract

In this paper, we propose a novel 3-D measurement system, which consists of 1) spatially phase-modulated light for encoding the projection angle, and 2) the three-phase correlation image sensor (3PCIS) for decoding amplitude and phase of time-varying intensity of light. The projection angle map is obtained based on spatio-temporal phase unwrapping method we propose. We implemented this system using our CMOS 00 times 200$ -pixel 3PCIS, and successfully reconstructed the range map in real time.

View full abstract

The model error is estimated by the inner function, which is implemented in to the controller. This scheme yields the new robust stability criterion, which does not depend on the model error. The error-free robust stabilization is proposed on the basis of the criteria.

View full abstract

The sufficient condition for robust stability of the approximate Smith predictor control is proved. It is shown that the stability area becomes wider in comparison with the original Smith predictor.

View full abstract

In this presentation, coefficients of given characteristic polynomial are polynomials of interval parameters. D-stability conditions will be given by combining the Jury-Pavlidis theorem with monotonicity. And approximated circle of D-stability region will be presented. Using stability conditions with monotonicity and the approximated circle, we will verify D-stability. The illustrative example will be given.

View full abstract

This paper deals with the mixed H₂ / D-stability control problem. In this paper, we give an iterative algorithm for finding a sub-optimal controller of this problem. The proposed algorithm produces a controller which satisfies a necessary condition to be globally optimal controller of the mixed H₂/ D-stability control problem. Numerical examples show the effectiveness of our algorithm.

View full abstract

An LP gas regulator vibrates abnormally due to deterioration caused by age. In this paper, we made a mathematical model of single-stage LP gas regulator in order to determine the cause of abnormal vibration and perform a computer simulation of the behavior of the vibration. We understood that the abnormal vibration is caused by unstable phenomena in the feedback system.

View full abstract

The health monitoring of propulsion systems is one challenge in space launch vehicles, particularly for manned space missions. Inverse problem-based strategy provides a new solution to the design of model-based fault diagnosis methods for monitoring the health of propulsion systems. This paper presents a decomposition approach to inverse problem-based fault diagnosis for a class of liquid rocket propulsion systems.

View full abstract

This article presents a method for setting the smoothing constant in the exponential smoothing. In a practical situation of the trend estimation, the smoothing constant is set relying on an analyst's background. We propose an adaptive setting method of the smoothing constant by solving the problem of tuning the gain of the Kalman filter.

View full abstract

In order to consider a failure diagnosis of fuel cells, we develop an experimental DMFC(Direct Methanol Fuel Cell), and provide some evaluations and analysis of the experimental results to make a mathematical model used for our diagnosis. By using experimental data, we identify the dynamics of the DMFC using the non-parametric identification method.

View full abstract

When a large number of input-output data are stored as database, local modeling and local controller design is very efficient since only the neighborhood information on the query point of the state space is required to construct the predicted output and design the output tracking controller. The proposed algorithm is applied to modeling of the MR damper and controller design for vibration reduction.

View full abstract

In this paper, we first derive nonholonomic dynamic systems with affine constraints under assumption of complete nonintegrability. We next show that the system is strongly locally accessible at any configuration point and small-time locally controllable under assumption of Sussmann's condition at any equilibrium. Finally we show the necessary condition of locally asymptotic stabilizability for the system at any equilibrium.

View full abstract

In this paper, we discuss a capability of a boundary control for disturbed distributed-parameter port-Hamiltonian systems. First, we show that a disturbance of the system can be decomposed into a boundary energy structure and a distributed energy structure uniquely. Next we define a boundary controllability of the system and we make the limitation of boundary control clear. Finally we give some examples to show the decomposition process.

View full abstract

This paper describes a robust stability analysis of discrete-time and discrete-value (discretized/quantized) nonlinear control systems in a frequency domain. By considering a threshold in the input of nonlinearity, a robust stability condition in a global sense is given. In this paper, a new technique for discretizing is proposed to relax the assumptions. Numerical examples are shown to verify the results.

View full abstract

We study L2 gain property for switched systems composed of continuous-time and discrete-time subsystems. When common Lyapunov function exists in L2 sense, we show the switched system has L2 gain less than gamma under arbitrary switching. When no common Lyapunov function exists, we use piecewise Lyapunov function showing the system achieves an ultimate L2 gain under average dwell time scheme.

View full abstract

Assigning instability to some singular points is useful for global asymptotic stabilization if a nonlinear control system is not globally asymptotically stabilizable to a single singular point. If the linear approximation is controllable, the Conley index of the singular point can be homotopy equivalent to one of hyperbolic singular point with any dimensional unstable manifold by using the discontinuous feedback.

View full abstract

Anti-roll tank is an effective device to decrease the ship's roll motion. An investigation of U-shape anti-roll tank was conducted with model experiment based on oscillation platform in this paper. The key problem realizing optimal control of tank model was solved very well, and the stabilizing effect of the designed tank model was validated by experimental results.

View full abstract

In a previous work, we proposed controllers that do not maintain the direction of the Jurdjevic-Quinn controller for a nonlinear system such that the hBcnorm of inputs is less than one. In this paper, we provide inverse optimal regulators that hold the directions of the proposed controllers.

View full abstract

We consider a particular type of LQ control problem, in which the control power is excluded from the cost function and state-observation noise is assumed. Our goal is a state feedback law that minimizes the cost function irrespectively of the output coefficient. A necessary and sufficient is found for a state feedback with constant gain to be optimal.

View full abstract

In this paper, an optimal feedback control law will be designed for linear building structures under seismic excitation. Using the proposed method, the vibration control of a singledegree-of-freedom structure with an active tuned mass damper (ATMD) will be studies. Using the analytical solutions, the optimal damping parameter for the auxiliary mass is also derived. Finally, the optimal displacement feedback control law is presented.

View full abstract

A Japanese government scientific research funding has been directed toward the promotion of an advanced plant production system involving intensive IT applications. Key technologies focused in the research are Speaking Plant Approach, soft-sensing, and WWW networking. An example of the speaking plant approach including soft-sensing of leaf surface temperature and Web based data handling will be discussed in this article.

View full abstract

In this tutorial, the outline of the nonlinear time series analysis scheme consisting of time delayed embedding technique, correlation dimension analysis, Lyapunov exponent analysis, nonlinear deterministic prediction etc. is introduced. As a part of tutorial session, I explain the general scheme of nonlinear time series analysis with a few case studies. Based on the scheme, I try to expand it to very short but ensemble ecological time series data set.

View full abstract

A neural network model which predicted plant shoot-tip temperature was constructed on six crops. The model had three layers, and the input was the environmental conditions of the greenhouse. A simulation study was performed with the constructed model, and it was found that the model was applicable as a tool to monitor plant shoot-tip temperature in commercial greenhouses.

View full abstract

An outline of the Discrete Element Method (DEM) which had been popular as a tool for solving dynamics of granular media is summarized in terms of its formulation. Some examples of DEM in terramechanics, where interactions between terrain and machine are the main subject of study, are introduced. Lastly, an attempt of DEM application to plant root-soil interaction is discussed.

View full abstract

A real-time machine vision system to guide a harvesting robotic manipulator for the red Fuji apples was developed. It is composed of a color CCD video camera to acquire Fuji apple images and a PC to process the images. The machine vision system was able to recognize and locate the fruit in less than one second.

View full abstract

This paper proposes a two-link segment model of a standing human controlled by reflex and voluntary movements. Time responses of body sway in physiological experiments are replicated by the simulation using the two-link model. Simulating a decline of human standing ability due to aging, we clarify the relation between the cause of falling and loss of function in old age.

View full abstract

This study investigates a neuromuscular system in reaching movements. A control mechanism for achieving antagonistic behavior is incorporated into a neuromuscular system and a performance criterion is defined from physiological viewpoints. Produced and observed trajectories are compared and significant features of the proposed neuromuscular system are evaluated. Moreover, characteristics of a neuromuscular system are investigated according to simulation results.

View full abstract

We compared EEG responses during various types of voluntary movements, i.e., self-paced and cue-guided, single and rhythmic, and actual and imaginary movements. We applied second-order statistics based blind source separation for de-noising.

View full abstract

A double series expansion technique was introduced to analyze translation and rotation of a spherical cap bio molecule in the stokes flow.

View full abstract

A generative methodology is proposed for the effective design of a product/project-oriented virtual enterprise. This methodology is distinguished into three stages: identification of business process, information modeling, and process planning & partner selection. Detailed analysis is focused on information modeling of product/project features and partners’ behavior & performance, to couple product/project requirements with partners core competencies.

View full abstract

We developed the object-oriented engineering tool for application programming of the controller based on the international standard IEC 61131-3 to improve software productivity. To realize continuity and translatability from conventional register-based programming, functions which complement and extend the international standard IEC 61131-3 were added by our unique and proper technologies. An engineering environment for easy software reusing was also realized.

View full abstract

This paper describes the architecture of a retargetable ladder diagram based software tool for programmable logic controllers (PLCs), which are based on the IEC 611311-3 specification. We propose here another approach of rapid prototyping process control with PLCs. The tool is designed in order to support as much as possible many embedded controller platforms or commercial PLCs.

View full abstract

It has been difficult to avoid falling over for walking robot with only legs. However, by adding arms, the robot is possible to transit from biped-walking to more stable quadruped-walking when the robot cannot maintain biped-walking. We demonstrate the limit of stable gait and that the robot can maintain locomotion through the transition to stable quadruped-walking despite applied external force.

View full abstract

This paper deals with an adaptive dynamical model for collective agent systems. The adaptation of collective behavior is modeled by the imitation dynamics, where the speed of collective strategy formation is affected by the group size. This paper describes a kind of inter-group dynamics, which adjusts group size such that the performance of collective behavior is improved.

View full abstract

The authors are studying on a simple scheme to control collective movements of autonomous mobile robots. In this article, a new manipulation method using a pre-compensator for its broadcast command is introduced. It enables to design dynamics of the center of the herd and that of each robot differently. Moreover, influences of an obstacle are analyzed. The validity is confirmed by simulation.

View full abstract

The dynamics of multi-link system can be represented and computed by the link dynamics and the mutual interaction among the links. Then, the static and dynamic characteristics of each link can be represented by only the information from itself and the adjacent link. In the present paper, both the inverse kinematics and dynamics of multi-link system are described by a distributed representation

View full abstract

This paper describes a testbed facility - the HoTDeC (HOvercraft Testbed for DEcentralized Control) - developed by the authors at the University of Illinois, consisting of multiple autonomous hovercraft vehicles which are wirelessly networked. This facility provides a flexible and state-of-the-art testbed for experimentation with inter-networked vehicles and sensors for decentralized and cooperative control in a dynamically nontrivial setting.

View full abstract

This paper describes possibility of robot assisted activity, RAA, and robot assisted therapy, RAT, through the fieldwork in the aged persons’ home and the pediatrics ward using several kinds of robots on the market.

View full abstract

The paper reports on a trial of Robot-assisted activity (RAA) for elderly people with Senile Dementia. The study subjects were eight elderly residents (68-89 year old) with Senile Dementia in a Group Home. The session was considered of approximately thirty minutes with the pet-type robot AIBO. The activities are useful to for smoothing communication with staff in a group home and establish friendly relations with occupants.

View full abstract

A long-term experiment of robot assisted activity for elderly people has been conducted at a health service facility for the aged since Aug. 2003. Two therapeutic seal robots, Paro, were introduced there. This paper describes an interim report of the experiment for three months.

View full abstract

Robot assisted activity (RAA) was carried out using various kind of pet-type robots for preschool and school age child patients in children’s hospital. RAA was carried out once a week at playroom in pediatrics. Change of mood of child patient is also investigated by administer a questionnaire using face scale after RAA.

View full abstract

This paper presents a new algorithm for global search using multi-points searches. Meta-heuristic algorithm has the layered structure which consists of global search operation and local search operation for every single point. The results of the local search are used as the initial conditions of the global search. We demonstrate that meta-heuristic algorithm can efficiently optimize various functions by applying meta-heuristic algorithm to some famous benchmark functions.

View full abstract

Cooperative collision avoidance for multiple mobile robot systems is a subject of this paper. Previously, the authors have proposed a series of the CCA methods based on a concept of the Common Velocity Obstacle. In this paper, the Iterative Decomposed CCA method is proposed, where residual of the existing DCCA method is utilized as a feedback input and reduced iteratively.

View full abstract

The present paper proposes a new type Fuzzy CSP that have a mixture of discrete and continuous domains, and Spread-Repair algorithm. We show that this model, called HDFCSP(Hybrid Domain Fuzzy CSP), can be solved by Spread-Repair, an extension of the well known iterative improvement algorithms.

View full abstract

A novel algorithm for amplitude and phase estimation of the multi-component signals is presented in this paper. The proposed algorithm is based on numerical differentiation and central Lagrange interpolation, and needs 1 signal sample with 7 points and 1 computation process. A testing example is given to illustrate the proposed algorithm, with the simulation is carried out in Matlab software.

View full abstract