This paper refers to Hardware Sortware Codesign from the view point of testing and verification of system LSI. First, we summarize problems regarding testing, verification and validation. Then, we consider a Hardware Sortware Codesign methodology maintaining its reliability. Finally, we introduce VCDS system developed by STARC.

View full abstract

This paper investigates some new trends in a control system based on Programmable Logic Control (PLC) from viewpoint of hybrid dynamical system. The way of handling some sophisticated control requirements, which are specified with taking into consider physical variables and time, is introduced. Also, the usefulness of these ideas are demonstrated by some practical applications.

View full abstract

In this paper, we review some recent results on supervisory control of partially observed discrete event systems. In order to solve a nonblocking supervisory control problem, we need to compute an L_m(G)-closed, controllable, and observable sublanguage of a given specification language. We present an algorithm for computing such a sublanguage which is larger than the supremal L_m(G)-closed, controllable, and normal sublanguage. We also consider a problem to find an observable event set with minimum cardinality which guarantees observability. The problem is computationally hard since its corresponding decision problem is NP-complete.

View full abstract

Action imitation, cooperation and communication share the same problem structure where two indepedent autonomous interaction dynamics should be coupled without destruction. We propose that this can be achieved by coupling via the “points of behavior articulation”. They play an important role both in action recognition and action exectution. Information extracted around the articulation points are invariant against perturbation and environmental fluctuations. Motion control applied at articulation points can achieve robust goal achievements while exploiting natural dynamics and tolerating fluctuations due to environmental conditions. The constraints which give rise to the articulation are largely dependent upon the body structure and dynamics. Maybe that is why humans share common articulation. Early experiments with a neural architecture for acquisition of the articulation structures are also presented.

View full abstract

In information communication networks, several control schemes are applied to node systems, circuits, and other equipment, and they interact with each other. So we can say that “information communications,” including telephony, computer communication, and the Internet, are made possible by these complicated control schemes. This paper describes several typical control schemes which are applied in this information communication. First, we classify networks into two-telephone network and IP network- and introduce typical control techniques in these networks. Finally, we also describe examples of congestion control schemes in telephony, VoIP services, and peer-to-peer communications.

View full abstract

what kinds of system are required aud how they can be constructed, as the basic industry promotion politics, for purpose that a new industry after another gets together in Kyoto to create an spontaneous industry accumulation with a keyword “IT”, based on the powerful potential of industry which is composed of what Kyoto has long attracted such as excellent companies, universities and colleges as well as many venture business companies and research institutes in Kansai Science City and Kyoto Research Park.

View full abstract

In this paper, an application of genetic algorithm for a quadrupedal robot is described. Each leg of the quadrupedal robot has a slider-crank mechanism and is driven by an actuator. To generate the optimum gait pattern for the robot, the quisi-optimum phase shift of the actuator in each leg is evolutionary generated by the genetic algorithm so that the walking distance increases within the trial period. The experimental results demonstrate that the proposed scheme is effective for generation of the optimum gait patterns and the robot can walk smoothly with the generated gait patterns under some walking conditions.

View full abstract

A number of design methods for output-feedback decentralized stabilizing controllers have been proposed for large-scale linear systems. In the centralized case, it is well known that this output-feedback synthesis problem is reduced to an optimization problem of linear matrix inequalities (LMI’s). In this paper, to design decentralized controllers, we restrict some variables of those LMI’s to block-diagonal forms. However, since this restriction is not enough to obtain decentralized controllers, we also introduce a certain replacement of variables. Since the latter causes the LMI’s to turn into bilinear matrix inequalities (BMI’s), we convert BMI’s into LMI’s by fixing some variables in the BMI’s. We get two kinds of LMI’s which are obtained by fixing some variables and solve those LMI’s alternately. The effectiveness of this design method is illustrated by numerical examples.

View full abstract

This paper studies a scheme for blind source deconvolution where the source signals are mixed by a dynamical system. Under the assumption that the mixing system belongs to a given class, a learning law is given in order to recover the original signals.

View full abstract

The stochastic models of evolution of phase boundaries are proposed by using the stochastic partial differential equations of Stratonovich type. Two types of evolution of phase boundaries, growth with constant speed and mean curvature flow, are considered in this paper. And the influence of disturbances to the behavior of such phase boundaries is studied through simulations.

View full abstract

In various fields of engineering, many kinds of patterns are observed including spiral and target patterns in BZ reaction. In this paper, focusing our attention to the pattern formation by Turing instabilities in stochastic reaction diffusion systems, the influence of disturbances to the pattern formation is studied through numerical simulations.

View full abstract

This paper addresses a smart flexible beam using a piezoelectric material with a self-maintenance for the characteristics variation. Perturbations of the piezoelectric materials capacitance of the flexible beam are caused by temperature fluctuation or breaking of wire. This paper proposed that the self-maintenance of a self-sensing system with a dynamic compensator can achieve to realize by adopting the system identification and the controller switching. Perturbation of the piezoelectric materials capacitance, experimental results by using the system identification and controller switching demonstrated the effectiveness of the proposed method.

View full abstract

The interplay between information science and control theory is discussed. In order to deal with a complex system in control theory, it is expected to be fruitful to import techniques from statistics, machine learning theory, information theory, and signal processing. It also seems promising to apply control theory to dynamical systems that appear in the mentioned fields of information science. Two results are presented as research in this direction. One is on a randomized algorithm for gain-scheduled control and the other is on non-stationary independent component analysis.

View full abstract

In the first half, a modified LQ performance index which has a parameter α is given, and a constant output feedback gain is obtained by minimizing it for a large α. In the last half, a necessary condition for optimality is obtained for the modified LQ performance index, and a homotopy approach is taken to obtain locally optimal solutions. But a numerical example shows that the solutions of the corresponding Riccati equation are not connected to the solutions.

View full abstract

This note analyses zeros of discrete-time models for mechanical systems expressed by matrix second-order equations and shows that no zeros of the discrete-time models reach the unit circle at any sampling period. Combining the above result with a sufficient condition for the zeros to be stable for sufficiently small sampling periods, it is shown that the zeros of discrete-time models for most of the mechanical systems are located inside the unit circle for all sampling periods.

View full abstract

This paper illustrates simulation of retrieval of a tethered satellite system. The simulation takes the mass and flexibility of the tether into consideration, and its model is called a flexible model. In the control design, the mass and flexibility of the tether are neglected, and the controlled system is expressed by a bilinear state equation, which is called a simple model. We compare simulation results of the flexible model and the simple model under nonlinear receding horizon control, which is designed for the simple model, and inspect the influence of the tether-flexibility.

View full abstract

It is known that for uncertain systems with delta-operator diamond characteristic polynomials, extreme point results do not hold, making their stability test a tough work. In this paper, we propose a stability analysis method for the system using directional stability radii which are capable of measuring stability margin for coefficient perturbations with restricted directions. The feature of the method is that the larger stability margin becomes, the easier the job is.

View full abstract

A new algorithm is proposed to compute complex stability radius for perturbed state space models using matrix sign function. The algorithm is motivated by H_∞ norm computation of transfer function matrices. The idea behind the proposed method is to combine the bisection method with a fast algorithm for H_∞ norm computation. Numerical examples are provided.

View full abstract

This report presents methods to design H_∞ and H₂ deconvolution filters for 2-D discrete systems. Since both methods are based on LMI, we can readily obtain such deconvolution filters which satisfy a given norm condition. An illustrative example is also studied.

View full abstract

This paper presents a method to design an optimal regulator for discrete-time systems, which consists of a set of controllers that stabilize a given plant when they act altogether, as well as when any one of them fails. It is shown that this problem can be solved by an iterative algorithm using LMI. The efficiency of the proposed method is illustrated through a numerical example.

View full abstract

Delayed feedback control (DFC) is well-known as a practical method for stabilizing unstable periodic orbits (UPOs) embedded in chaotic attractors. The stabilization of magneto-elastic system applied to DFC has been achieved experimentally. However, the mechanism of stabilization has not been clearly understood, since the system has been analytically discussed under restriction within the neighborhood of a stabilized UPO. In this paper, coexisting orbits with the stabilized UPO are numerically discussed. The existence of solutions approaching UPO along unstable coexisting orbit is shown.

View full abstract

In this paper, we study the relationship between observability and normality arising in supervisory control of discrete event systems under partial observation. We prove that observability together with a condition on unobservable events is equivalent to normality in the presence of controllability. This result means that a supervisor synthesized for the supremal controllable and normal sublanguage of a specification language is a maximally permissive one under a restriction that no unobservable event is disabled.

View full abstract

This paper evaluates control methodologies for systems with constraints, utilizing a laboratory equipment for controls of a yawing motion of a RC helicopter. The methodologies implemented here are static and dynamic anti-windup compensator and reference governor. And from experimental results the effectiveness of them is compared and evaluated.

View full abstract

This paper is concerned with iterative learning control of Hamiltonian systems, which is applicable to electro-mechanical systems. We propose a new learning control scheme based on the self-adjoint structure of the variational of those systems. This method dose not require precise information about target systems such as the physical parameters of those systems. And we demonstrate the effectiveness of the proposed method by experiments.

View full abstract

In this paper, we propose a discontinuous feedback controller for a two-wheeled mobile robot which moves in the presence of obstacles. The controller is designed by utilizing a positive definite function (Lyapunov function) which is minimized at a desired point, and it is possible to change the trajectory of the controlled system according to the choice of the Lyapunov function. In this paper, we consider some simple obstacles such as a flat wall, and construct a Lyapunov function corresponding to them. The controller derived from the Lyapunov function makes the system converge to the desired point avoiding the obstacles.

View full abstract

In this paper, we study the solvability of optimal singular control problems by making the most of the generalized Legendre-Clebsch condition for the optimality of the singular control. Using the Hamiltonian expression of Maximum Principle and the variational formula of Rozonoér, we discuss the local attainability along the singular trajectory and the conditions of invariance, in other words, the local strong attainability distribution. It seems possible to find out solvable cases of the optimal singular control problems left unsolved since the sixties of the 20^th century.

View full abstract

A common quadratic Lyapunov function guarantees asymptotic stability of a set of systems. The common Lyapunov function problem to be study here, is to specify a set of systems which have a common quadratic Lyapunov function guaranteeing asymptotic stability of its every member system. Several sufficient condition of linear systems which have a common quadratic Lyapunov function, are already known. Furthermore, there exist some parallel results between for continuous-time and discrete-time linear systems. In this paper, we add a new sufficient condition of discrete-time systems with this property using diagonally dominant matrices. This condition provides a counterpart of a known result for continuouse-time systems.

View full abstract

This report introduces application results of QE method to common infinity (l_∞) Lyapunov function problem for continuous-time linear time-invariant systems. First, the report presents an application of QE method to find the existence condition of a common l_∞ Lyapunov function for a set of continuous-time linear time-invariant systems and gives a numerical example of a pair of second order systems. Next, by giving numerical examples, the report makes clear about the non-inclusion and non-exclusion relation between the existence conditions of common l_∞and l₂ Lyapunov functions for a pair of second order continuous-time linear time-invariant systems.

View full abstract

The common Lyapunov function problem is a problem which examines existence conditions of a common Lyapunov function for several linear time-invariants systems. The problem arises, for example, when we consider the stability of a switching system which consists of several linear time-invariant systems. Although the problem for the existence of a quadratic Lyapunov function has been discussed, l_∞ Lyapunov function counterpart seems to have not been treated. In this study, we examine the common l_∞ Lyapunov function problem for both continuous-time and discrete-time systems. For both cases conditions for some subsets of the set of systems which have a common l_∞ Lyapunov function were clarified.

View full abstract

Strictly positive real control problem is considered for a linear time-invariant plant, where an internally stabilizing controller is designed such that the resultant closed-loop transfer function is strictly positive real. The solvability condition of this problem includes not only inequality constraints but also equality constraints. This paper presents a necessary and sufficient condition for the existence of a controller in terms of strict linear matrix inequalities.

View full abstract

The purpose of this research is to propose a method to evaluate the travel time savings by introducing intelligent transport systems (ITS). For the purpose we employed the computer simulation which is based on the cellular automaton method.

View full abstract

The purpose of this paper is to evaluate the value of a manufacturing order where the market price of a product is fluctuating randomly. Applying the Black-Scholes equation, we show the replicating assets of the manufacturing order. In addition we show how to simulate the market price of a product then propose the analysis method which is based on the value at risk(VaR).

View full abstract

This paper considers a method of fault detection of boundaries of plates by using segmented piezoelectric films under sinusoidal excitation. Boundaries are modeled by springs and one of them varies from healthy state. To detect the broken boundary, we examine minimum values of a cost function of springs at every boundaries and regard the boundary which gives minimum cost as the broken one. Owing to the preferable properties of piezoelectric film, it is expected that it can be applied to structures easily and has high sensitivity to the changes of the boundary condition. Numerical examples are shown to illustrate the efficiency.

View full abstract

In an intravenous anesthesia, the Bispectral Index (BIS) is a reliable index to anesthetic depth. We developed an intravenous anesthesia control system using a model predictive controller based on BIS. Also, we developed an on-line identification method of pharmacodynamics and delay time for each patient, because the individual differences of the responses of BIS to anesthetic drug are very large. However, it is not easy to design the model predictive controller so that specified amounts of delay and gain margins are guaranteed. In this paper, an intravenous anesthesia control system using a state-predictive controller, for which a theoretical method is available to assure delay and gain margins, is developed.

View full abstract

During surgical operation, patients must be kept under appropriate sedation. In order to keep patients under appropriate sedation, we developed an intravenous anesthesia control system using a model predictive controller, which uses propofol as the anesthetic drug and Bispectral Index (BIS) as the index of sedation, and has an identification function of the individual response of each patient. To make clinical applications, we have added a fail-safe function to this system for preventing drug over-infusion or intra-operative arousal. With the permission by the ethical committee of Kyoto University Hospital, 12 clinical applications were made from December 2001. In this paper, we report the fail-safe function and the result of clinical applications.

View full abstract

This paper presents a method for human-computer cooperative control of a mobile robot, where the system supports some cooperative action between manual and automatic modes. The proposed method uses variable gains which change the connection of dual modes. The robot is usually controlled by a computer using known information, but if it is necessary for human operator to adjust the motion of the robot, human operator can intervene in computer operation to reduce the speed of the robot and change the course of the motion by moving the joystick. The method, proven by some actual trials in a PWS mobile robot, realizes efficiencies in the operation of mobile robot with simplicities of commands and flexibility of human operator.

View full abstract

This paper tries to explain a mathematical foundation of neural network learning. Hierarchical learning machines such as neural networks and gaussian mixtures are non-identifiable learning machines, resulting that the conventional statistical asymptotic theory can not be applied. For such learning machines, the Bayes estimation is more appropriate than the maximum likelihood method. We show the reason why neural networks in Bayesian estimation is useful in practical applications.

View full abstract

The author proposes “Telemetric Rescue Dummy (TRD)”, which can simulate a victim for evaluating rescue equipments and rescue robots. In various fields, dummies that simulate human bodies have been developed until now. By contrast to them, functions required for simulating rescue tasks are discussed. Then the concept of telemetric rescue dummy is clarified. Moreover, DAMIYAN, dummy developed for the 1st Rescue Robot Contest, is introduced as a concrete example of the TRD.

View full abstract

The rescue robot contest is a robot contest with the society. The relation to the cooperation of the contest and industry, government and university is shown from a new viewpoint. As a result, it is described that it is very significant to affect the contest.

View full abstract

Recently, the problem of estimating motion fields from image sequences is becoming important for root vision and so on. We already proposed a method for estimating an entire motion-vector field from a given set of image-sequence data. In the method two different methods, the matching method and the gradient method are realized and fused by the use of learning and function approximation abilities of neural networks. The purpose of this study is to investigate the performance of the proposed method through experiments in several situations. It is shown that the proposed fusion is effective and makes it possible to estimate motion fields more accurately.

View full abstract

The effects of the external noise and reverberation time of the acoustical propagation space on two psychological impressions for the noise/audio signal and the listening score, in the case of listening to monosyllable/word audio signal, is discussed. Cncretely, the two psychological impressions and the listening score are considered in condition with the speech interference level and the revervation time. Next, the prediction problems of the psychological impressions and the listening score are considered.

View full abstract

In this paper, we propose a new segmentation method for brain blood vessels of MR angiograms. This method consists of the following processes: create three (x,y,z-axis) MIPs, perform region growing on each MIP, and integrate these results in a 3D space. Applying this method, which has both features of MIP and region growing, we can extract 3D blood vessels with accuracy and reliability.

View full abstract

This paper deals with a method of superimposing a coronary arterial tree on a myocardial SPECT image. We can grasp the myocardial function more easily in connection with the shape of the coronary artery. First, we estimate a rough shape of the left ventricle (LV) by using SPECT data and project it on a pair of bi-plane CAG images. Then we obtain two 2-D coronary images registered to the surface of the LV model. By maximizing a matching degree, registration between CAG images and SPECT image is performed automatically. We can obtain 2-D superimposed image on bull’s eye map. and 3-D image on LV model simplily.

View full abstract

Recently introduced multi detector CT(MDCT) allows us to aquire dynamic 3D images(4D images) of the heart with high spatial and temporal resolution. For the analysis of cardiac function from 4DCT data, automatic segmentation of the heart, auto extraction of LV in particular, is important. In this paper, we presented an automatic method of extraction from 4DCT data by following automatic reconstruction of short axis slices.

View full abstract

We investigated characteristics of several matching measures between preoperative 3D-CTA and intraoperative fluoroscopic images. In this study, DRR (Digitally Reconstructed Radiograph) are generated by voxel projection of 3D-CTA after extracting an aorta region. By increasing/decreasing CT value in the aorta region of CTA, DRR with/without contrast media injection are obtained. Subsequently we calculate matching measures between DRR and fluoroscopic images iteratively by changing imaging parameters. The most similar DRR to fluoroscopic image is selected. From simulation results, we think mutual information and residual of M-estimator are suitable because estimation error is small and distribution of matching measure has sharp peak.

View full abstract

X-ray mammography (mammogram) currently is a common imaging method used to check primarily the presence of breast pathologies. Because images are densely textured with low signal to noise ratio, it may happen that diagnosis of small pathologies, such as cancers, is missed. In this work we propose a robust fuzzy filtering for image enhancement and noise reduction at the same time. Gray scale intensity is considered an element of a fuzzy set and is associated to gaussian membership functions. The output of fuzzy filter is a mapping of target pixel to a new value that may correspond to background object, pathological calcification, or fibrous tissue. Our filter, eliminating high frequency noise, assures that micro-calcifications are well enhanced.

View full abstract

This paper assumes that the inconsistency of the pairwise comparison in the AHP methodology comes from the decision maker’s distortion over his preference transitivity. We propose the method to modify the weights of the criterion where the pairwise comparison among alternatives is inconsistent. Our method undervalues the criteria under which preference of the decision maker among alternatives is inconsistent, and highly evaluates the criteria where his preference is clear.

View full abstract

Recently, Triantaphyllou has reported that, when evaluating binomial relation by the original analytic hierarchy process (AHP), the ranking is different from that obtained when evaluating all alternative simultaneously, even if the preference transitivity holds and if all pairwise comparison matrices are completely consistent. In this paper, we show that the aspiration level AHP, one of modifications of AHP proposed by Tamura et al., does not cause such irrational ranking for completely consistent case. We also estimate how often irrational ranking occurs in practical case by experimental analysis.

View full abstract

In this paper, we treat goal programming problems with interactive fuzzy coefficients. To treat the interactivity between fuzzy numbers, we introduce an oblique fuzzy vector. Goal programming problems with oblique fuzzy vectors are formulated based on the fractile optimization model using a necessity measure in possibilistic linear programming. It is shown that the problem is reduced to linear programming problems with a special structure. Utlizing the special structure, we propose an algorithm based on Benders decomposition method.

View full abstract

Linear programming problems are widely used. The problems such as optimal operating problems have block angular structure. It is expected that the block iterative method can reduce the computational time by utilizing the typical structure. In this paper, a new method which utilizes the predictor-corrector method and the block iterative method is proposed. By numerical experiments, it is shown that the new method is effective and that the computational time of the new method is reduced to about half of the conventional method.

View full abstract