제어로봇시스템학회:학술대회논문집
Institute of Control, Robotics and Systems (ICROS)
 기타
Domain
 Machinery ＞ Robot/Automated Machinery
1995.10

Generalized predictive control based on the parametrization of twodegreeoffreedom control systemsWe propose a new design method for a generalized predictive control (GPC) system based on the parametrization of twodegreeof freedom control systems. The objective is to design the GPC system which guarantees the stability of the control system for a perturbed plant. The design procedure of our proposed method consists of three steps. First, we design a basic controller for a nominal plant using the LQG method and parametrize a whole control system. Next, we identify the deviation between the perturbed plant and the nominal one using a closedloop identification method and design a free parameter of parametrization to stabilize the closedloop system. Finally, we design a feedforward controller so as to incorporate GPC technique into our controller structure. A numerical example is presented to show the effectiveness of our proposed method.

In many cases of robust stability problems, the characteristic polynomial has real coefficients which or nonlinear functions of uncertain parameters. For this set of polynomials, a new stability easily checking algorithm for reducing the conservatism of the stability bound are given. It is the new stability theorem to determine the stability region just in parameter space. Illustrative example show that the presented method has larger stability bound in uncertain parameter space than others.

This paper considers the linearquadratic optimal regulator problem for nonstandard singularly perturbed systems making use of the recursive technique. We first derive a generalized Riccati differential equation by the HamiltonJacobi equation. In order to obtain the feedback gain, we must solve the generalized algebraic Riccati equation. Using the recursive technique, we show that the solution of the generalized algebraic Riccati equation converges with the rate of convergence of O(.epsilon.). The existence of a bounded solution of error term can be proved by the implicit function theorem. It is enough to show that the corresponding Jacobian matrix is nonsingular at .epsilon. = 0. As a result, the solution of optimal regulator problem for nonstandard singularly perturbed systems can be obtained with an accuracy of O(.epsilon.
$^{k}$ ). The proposed technique represents a significant improvement since the existing method for the standard singularly perturbed systems can not be applied to the nonstandard singularly perturbed systems. 
Various methods and techniques have been proposed for solving optimization problems; the methods have been applied to various practical problems. However the methods have demerits. The demerits which should be covered are, for example, falling into local minima, or, a slow convergence speed to optimal points. In this paper, Likelihood Search Method (L.S.M.) is proposed for searching for a global optimum systematically and effectively in a single framework, which is not a combination of different methods. The L.S.M. is a sort of a random search method (R.S.M.) and thus can get out of local minima. However exploitation of gradient information makes the L.S.M. superior in convergence speed to the commonly used R.S.M..

In this paper we describe a new method for multimodal function optimization using genetic algorithms(GAs). We propose adaptation rules for GA parameters such as population size, crossover probability and mutation probability. In the self organizing genetic algorithm(SOGA), SOGA parameters change according to the adaptation rules. Thus, we do not have to set the parameters manually. We discuss about SOGA and those of other approaches for adapting operator probabilities in GAs. The validity of the proposed algorithm will be verified in a simulation example of system identification.

This paper shows adaptive control using RHPC(Receding Horizon Predictive Control) with equality constraint which applied to Electric Furnace. The control strategy includes monotonic weighting (improving transient response) and prefiltering (enhancing robustness), which is effective on real process. We can observe the performance of RHPC and confirm the practical aspect of RHPC with unmodelled dynamics through the experiment of Electric Furnace. Finally, this paper verifies the feasibility of RHPC to real process.

In this study, the fuzzy approximator and sliding mode control (SMC) scheme are considered. An adaptive sliding mode control is proposed based on the SMC theory. This proposed control scheme is that a adaptive law is utilized to approximate the unknown function f by fuzzy logic system in designing the sliding mode controller for the nonlinear system. In order to reduce the approximation errors, the differences of nonlinear function and fuzzy approximator, an adaptive law is also intoduced and the stability of proposed control scheme are proven with simple adaptive law and roburst adaptive law. This proposed control scheme is applied to a single link robot arm.

The automated chemical analysis shop floor are developed for the environmental pollution problems in our chemical analysis center. This shop floor have the several equipments include weight, pour, dry, heater, boiler, mixture, spectroscopy etc. And the material handling components are made up by the stored stack, conveyore, turntables, robot etc. Computer simulation has been an important tool for these complete design problem. We have designed the arangement of chemical equipments and material flow systems by using the simulator "AutoModII". "AutoMoII" is one of the advanced simulator, CADlike drawing tools with a powerful, engineering oriented language to model control logic and material flow. The result is the modeling of the chemical analysis system in accurate, three dimensional detail. We could designed the set able layout and scheduling system by using the AutoMoII simulator. AutoMoII simulator.

We proposed in this paper a systematic way for analyzing discrete event dynamic systems to classify faults and failures quantitatively and to find tolerable fault event sequences embedded in the system. An automated failure diagnosis scheme with respect to the nominal normal operating event sequences and the supervisory control problem for tolerable fault event sequences is presented. Moreover the supervisor failure diagnosis problem with respect to the tolerable fault event sequences is considered. Finally, a plasma etching system example is presented.

A new methodology for representing the interaction between machines and the interlock signals required in FMCs has been developed. ObjectOriented Philosophies (OOPs) lend themselves to the development of such a scheme. A methodology developed here regards the tasks to be performed by the cell or any of its constituent machines for being primal. Sensory signals indicating the changes of state pf machines are used to trigger or initiate tasks. A task may be simple and require a relatively short time to execute, or may be complex and lengthy. This methodology may be depicated by a set of diagrams called Task Initiation Diagram (TID) and their accompanying rules.

Adaptive algorithms based on gradient adaptation have been extensively investigated and successfully jointed with active noise/vibration control applications. The FilteredX LMS algorithm became one of the basic feedforward algorithms in such applications, but still is not fully understood. The error path model effect on the FilteredX LMS algorithm has been under the investigation and some useful properties related stability has been discovered. We are interested in utilizing the fact that the model error caused by the way optimizing the error path model in a view point of convergence speed of FilteredX LMS algorithm for pure tone noise suppression application without any performance loss at steady state.

This paper concerns an adaptive control scheme which is an extension of the simplified adaptive control. Originally, the SAC approach was developed based on the command generator tracker (CGT) theory for perfect model tracking. An attractive point of the SAC is that a control input can be synthesized without any prior knowledge about plant structure. However, a feedforward dynamic compensator of the CGT is removed from the basic structure of the SAC. This deletion of the compensator makes perfect model tracking difficult against even a step input. In this paper, an adaptive control system is redesigned to achieve perfect model tracking for as long as possible by reviving the dynamic compensator of the CGT. The proposed method is applied to slewing control of a flexible space structure and compared to the SAC responses.

An adaptive unified predictive control (UPC) algorithm is applied to a batch polymerization reactor for poly(methyl methancrylate) (PMMA) and the effects of controller parameters are investigated. Computational studies are performed for a batch polymerization system model developed in this study. A transfer function in parametric form is estimated by recursive least squares (RLS) method, and the UPC algorithm is implemented to control the reactor temperature on the basis of this transfer function. The adaptive unified predictive controller shows a better performance than the PID controller for tracking set point changes, especially in the latter part of reaction course when gel effect becomes significant. Various performance can be acquired by selecting adequate values for parameters of the adaptive unified predictive controller; in other words, the optimal set of parameters exists for a given set of reaction conditions and control objective.

In this paper, we propose a new rubust design scheme of a variable structure type model reference control (VSMRC) which can be applied to linear timevaring plants. Our idea is started from the hypothesis that the plant consists of two parts, i.e., one has timeinvariant parameters and the other has timevarying parameters. We consider the former the nominal part of the plant and the latter a kind of disturbance to the nominal one. In this design scheme, the ordinary VSMRC is adopted to the nominal part and the signum function is introduced to eliminate the influence of the disturbance.

The conventional neural network models are a parody of biological neural structures, and have very slow learning. In order to emulate some dynamic functions, such as learning and adaption, and to better reflect the dynamics of biological neurons, M.M. Gupta and D.H. Rao have developed a 'dynamic neural model'(DNU). Proposed neural unit model is to introduce some dynamics to the neuron transfer function, such that the neuron activity depends on internal states. Integrating an dynamic elementry processor within the neuron allows the neuron to act dynamic response Numerical examples are presented for a model system. Those case studies showed that the proposed DNU is so useful in practical sense.

This paper describes trajectory generation of a riobot arm by selforganizing neural networks. These neural networks are based on competitive learning without a teacher and this algorithm which is suitable for problems in which solutions as teaching signal cannot be definede.g. inverse dynamics analysisis adopted to the trajectory generation problem of a robot arm. Utility of unsupervised learning algorithm is confirmed by applying the approximated solution of each joint calculated through learning to an actual robot arm in giving the experiment of tracking for reference trajectory.

This paper considers the applications of cellular automata in order to design selforganizing artificial neural decision systems such as selforganizing neurocomputer circuit, machines, and artifical life VLSI circuits for controlling mechanical systems. We consider the Lsystem and show the results of growth of plants in artificial life.

In this paper, a new PID fuzzy controller(FC) based on parallel operation of PI and PD fuzzy control is presented. First, two fuzzy rule bases are constructed by separating the linguistic control rule for PID FC into two parts : one is e.DELTA.e part, and the other is .DELTAL.
$^{2}$ e.DELTA.e part. And then two FCs employing these rule bases indivisually are synthesized and run in parallel. The incremental control input is determined by taking weighted mean of the outputs of two FCs. The proposed PID FC improves the transient response of the system and gives better performance than the conventional PI FC. 
In this paper it is concerned to develop control method using jartest results in order to predict the optimum dosage of coaglant, PAC(PoliAluminum Chloride). Considering the relations with the reactions with the reaction of coagulation and flocculation, the five independent variables ( e, g, turbidity of raw water, water turbidity in flocculators, temperature, pH, and alkalynity) are selected out of parameters and they are put into calculation to develop a neural network model for PAC dosing process in water purification system. This model is utilized to predict optimum dosage of PAC. That is, the optimum dosage of PAC is searched in neural network model for PAC dosing process to minimize the water turbidity in flocculators. This searching is implemented by means of expert heuristics. The efficacy of the proposed contorl schemem and feasibility of acquired neural network model for PAC dosing contorl in water purification system is evaluated by means of computer simulation.

The problem of temperature control for rapid thermal processing (RTP) in semiconductor manufacturing is discussed in this paper. Among sub=micron technologies for VLSI devices, reducing the junction depth of doped region is of great importance. This paper investigates existing methods for manufacturing wafers, focusing on the RPT which is considered to be good for formation of shallow junctions and performs the wafer fabrication operation in a single chamber of annealing, oxidation, chemical vapor deposition, etc., within a few minutes. In RTP for semiconductor manufacturing, accurate and uniform control of the wafer temperature is essential. In this paper, a robustr controller is designed using a recently developed optimization technique. The controller designed is then tested via computer simulation and compared with the other results.

This paper proposes an autotuning method of feedforward signal in boiler control of thermal power plants by using the neural network. The neural network produces an optimal feedforward signal by tuning the weights of the network. The weights are adapted effectively by using the teaching signal of PI control output. The proposed method was evaluated based on a detailed simulator which expressed nonlinear characteristics of the 600 MW actual thermal power plant at load chaning operations, showed effectiveness in the learning of the weights of the neural network, and gave an accurate control performance in the temperature control of the system. Through the evaluation, the proposed method was proved to be effectively applicable to the actual thermal plants as the automatic adjustment tool.

When the ultrasonic transducer operating at l MHz for use in cleaning semiconductor wafers or other industsrial materials is driven from the MOSFET DCto RF inverter, the output power severely depends on the frequency of operation since the quality factor of the transducer is high. In order to tune to the eresonating frequency of the ultrasonic transducer, the drive signal frequency of the MOSFET power inverter is automatically scananed until the frequency is set at the resonating frequency of the ultrasonic transducer is maximized. The control circuit consists of an output power sensing circuit, a PLL controller, a frequency standard, and other peripheral circuits. The operation was satisfactory when the transducer having an output of 600 W at 1 MHz was used.

A newly developed discretetime control design method for impact machines is proposed. It is composed of identification and control using neural networks, where the optimal controller with saturationn and no use of velocity measurements is obtained. By computer simulation, the proposed method is demonstrated to be effective: as the training progresses, the cost function becomes smaller, the proposed control is superior to PID control tuned with ZieglerNichols (ZN) parameters; robust performance with respect to uncertainty, disturbances and working time is so good.

In this paper, an unknowninput proportional integral (PI) observer is presented and its applicability to the design of exact loop transfer recovery (Exact LTR) is shown. First, a sufficient condition for the PI observer to estimate the states of systems without knowledge of unknown input is derived. A simple existence condition of the observer is given. Under the conditions, the Exact LTR with specified observer's poles is achieved by the unknowninput PI observer.

In this paper, we consider the stabilization problem of nonstandard singularly perturbed systems by using state feedback. Different fro the existing sequenetial designn procedures, we propose a parallel design method to construct the stabilizing controller. The method involves solving two completely independent algebraic Riccati equations.

In this paper, though Simith controller is also used, we propose a new system configuration which can be regarded an SISO continuous nthorder plant with time delay of ktimes of a sampling period as a linear discrete (n + k)th order plant of which all state variables can be available. Consequently, we can pply linear control system design techniques which do not consider the existence of time delay to the proposed system.

First, in this paper we propose a new dead best control system design technique by which we can specify a transient response before the settling time. Though the resultant system has the same system configuration as Reference[1], that is realized by adapting the performance index which includes the term of the square of difference between specified and pracitical responses. Next, we state a technique which gives the dead beat control system robustness and construct a robust dead beat control system. Simulations of the proposed dead beat control and robust dead beat control systems show expected results.

This paper presents the onboard attitude determination algorithm for LEO (Low Earth Orbit) threeaxis stabilized spacecraft. Two advanced star trackers and a threeaxis Inertial Reference Unit (IRU) are assumed to be attitude sensors. The gyro in the IRU provides a direct measurement of the attitude rates. However, the attitude estimation error increases with time due to the gyro drift and noise. An update filter with measurements of star trackers and/or sun sensor is designed to update these gyro drift bias and to compensate the attitude error. Kalman Filter is adapted for the onboard update filter algorithm. Simulation results will be presented to investigate the attitude pointing performance.

This paper present an attitude control using quaternions as feedback attitude errors. The Euler's eigenaxis rotation provides the shortest angular path between two attitudes. This eigenaxis rotation can be achieved by using quaternions since quaternions are related with the eigenaxis. The suggested controller uses error quaternions and body angular rates and generates a decoupling control torque that counteracts the natural gyroscopic coupling torque. The momentum dumping strategy using the earth magnetic field is also applied in this paper to unload the angular momentum of the reaction wheels used in the attitude control.

Neural network has found wide applications in the system identification, modeling, and realization based on its function approximation capability. THe system governe dby nonlinear dynamics is hard to be identified by the neural network because there exist following difficulties. FIrst, the training samples obtained by the stae trajectory are apt to be nonuniform over the region of interest. Second, the system may becomje unstable while attempting to obtain the samples. This paper deals with these problems in discretetime system and suggest effective solutions which provide stability and uniform sampliing by the virtue of robust control theory and heuristic algorithms.

Characteristics of control system design using Universal Learning Network (U.L.N.) are that a system to be controlled and a controller are both constructed by U.L.N. and that the controller is best tuned through learning. U.L.N has the same generalization ability as N.N.. So the controller constructed by U.L.N. is able to control the system in a favorable way under the condition different from the condition of the control system in learning stage. But stability can not be realized sufficiently. In this paper, we propose a robust control method using U.L.N. and second order derivatives of U.L.N.. The proposed method can realize better performance and robustness than the commonly used Neural Network. Robust control considered here is defined as follows. Even though initial values of node outputs change from those in learning, the control system is able to reduce its influence to other node outputs and can control the system in a preferable way as in the case of no variation. In order to realize such robust control, a new term concerning the variation is added to a usual criterion function. And parameter variables are adjusted so as to minimize the above mentioned criterion function using the second order derivatives of criterion function with respect to the parameters. Finally it is shown that the controller constricted by the proposed method works in an effective way through a simulation study of a nonlinear crane system.

Fuzzy controllers have proven to be powerful in controlling dynamic processes where mathematical models are unknown or intractable and illdefined. The way of improving the performance of a fuzzy controller is based on making up rules, constructing membership functions, selecting a defuzzification method and adjusting inputoutput scaling factors. But there are many difficulties in tuning those to optimize a fuzzy controller. So, in this paper, we propose the lookup table based selforgenizing fuzzy controller (LSOFC) which optimizes lookup values resulting from the above fuzzy processes. We use the plusminus tuning method(PMTM), scanning the value through the processes of addition and subtraction. Simulation results demonstrate that the performance of LSOFC is far better than that of a nontuning fuzzy controller.

We developed a visual system that is able to track the moving objects within a certain range of errors. The visual system is driven by two DC servo motors that are controlled by a computer based on the visual data obtained from a CCD video camera. The software to track the moving objects is developed based on the PWM of the DC motors. Also, the problems how to implement a fuzzy logic control method and a neural network in this system, are also considered in order to check the control performance of tracking. The fuzzy logic algorithm is a powerful control technique for nonlinear dynamical system and also the neural network could be implemented in this system. In this paper, we present configuration of tracking system developed in our laboratory, the control methods of the visual system and the experimental results are shown.

A new design method is proposed for a fuzzy PD controller. By analyzing phase plane characteristics we can build and optimize the rule base of fuzzy logic controller. Also, a new gain tuning method is used to improve performance in the transient and steady state. The improved performance of the new methodology is shown by an application to the design of control system with a highly nonlinear actuator.

This paper describes a 3dimensional measuring method for cylindrical huge ingot pressed with forging machine. Target ingot we consider here is rotated around a fixed axis during measurement. Using an image processing technique every profile of crosssection is obtained, and 3D image is reconstructed. One method to calibrate the system setting is also presented. Experimental results reveal that the method are applicable and the algorithm is feasible.

It is important from prevention of the malfunction and an important accident by the failuer, to detect a failuer in revolution devices. The acoustic emission(AE) method is expected as means that defects an abnormal phenomenon of revolution devices earlily and utilized. Although a research example by the AE method is reported regarding a gears, little reserch has been conducted using the AE method for running gears in a bending fatigue process of spur gear teeth. Therefore, in this report, with two micro AE sensors attached to the side of the gear, AE was measured in a bending fatigue process of a carburizing gear by using the power circuratingtype machine and AE source location in gear teeth were required. By various analysis in these data, the AE characteristics in the fatigue damaging process of the gear tooth were determined.

A new method for grouping of relevant and equivalent inputs of a logical circuit was proposed by the authors by making use of pseudorandom Msequence correlation. The authors show in this paper that it is possible to estimate the input grouping from a part of correlation functions when we admit small percentage of error, whereas it is impossible to reduce the data necessary to estimate the grouping by use of the truth table method. For example in case of 30input logic circuit, the number of correlation functions necessary to calculate can be reducible from 1.07 * 10
$^{9}$ to 465. 
In this paper the attitude control system is developed for longitudinal motion of FoilCatamaran in regular waves with allmovable foils which attached to fore and after part of the ship and verified the system by theoretical calculation and modeltests. The linearized equations of motion of the ship is employed to apply the linear control theories, the PID control and the LQR. The strip method was used to calculate hydrodynamic coefficients and wave exciting forces of the demi hull, and unsteady hydrodynamic forces of foils are considered by using the result of Wu(1972). About 4060% of motions is reduced in experiments. The control system described in this paper is able to extended to 6DOF motions or control in irregular wave with trivial modification. And it is applicable to hull shape development for better seakeeping performance and to determine the size and the position of hydrofoils for the attitude control.

A magnetic levitation control system is nonlinear and very unstable. Thus there should be a stabilizing compensation network and a feedback path. Due to the levitation control a noncontact photoresistor sensor is generally used. One photocell provides a certain amount of variation in length by the ball shadow casted on the cell surface. Furthermore at the boundary of the cell, the linearity of sensitivity deteriorates severely. To overcome the constraints of the length and linearity, an efficient sensor array is deviced and applied in the feedback path of a largegap magnetic levitation control system. A number of CdS photocells and a summing circuit of the sensor output signals are used for a sensor array. The levitation length of a ball and the transient performances are main objectives of the largegap suspension system using the sensor array.

Shinohara, Shigenobu;Pan, Derong;Kosaka, Nozomu;Ikeda, Hiroaki;Yoshida, Hirofumi;Sumi, Masao 158
A new laser diode rangefinding speedometer is proposed, which is modulated by a pair of positive and negative triangular pulse current superimposed on a dc current. Since a target velocity is directly obtained form a pure Doppler beat frequency measured during the nonmodulation period, the new sensor is free from the difficulties due to the critical velocity encountered in the previous sensor. Furthermore, the different amplitude of the two triangular pluses are so adjusted that the measurable range using only one laser head is greatly expanded to 10cm through 150cm, which is about two times that of the previous sensor. The measurement accuracy for velocity of .+.6mm/s through .+.20mm/s and for range is about 1%, and 2%, respectively. Because the new sensor can be operated automatically using a microcomputer, it will be useful for application of a 3D range image measurement of a slowly moving object. 
For an accurate online measurement of the ship's attitude the paper develops an intelligent sensing system which uses one servotype accelerometer and two servotype inclinometers appropriately located on the ship. By considering the dynamics of the servocontrolled rigid pendulums of the inclinometers, linear equations for the rolling and pitching of the ship are derived separately from each other. Moreover, one accelerometer is used for extracting the heaving signal. Through the introduction of linear dynamic models and the linear observation equations for the heaving, rolling and pitching, the online measurement of the three signals can be reduced to the state estimation of the linear dynamic systems. A bank of Kalman filters is adaptively used to achieve the online accurate state estimation and to overcome changes in parameters in the linear dynamic models.

An intelligent system which is an integration of multifunctional instrumentation (MFI) and a neural network is discussed. According to some experiments of temperature and wind velocity it is clear that this system can learn the data structure of two parameters above. So it makes good performances for estimations of nonsample data.

Shinohara, Shigenobu;Hara, Katsuhiko;Toyoshima, Morio;Ikeda, Hiroaki;Yoshida, Hirofumi;Sumi, Masao 170
Recently, we proposed a compact digital vibrometer using a selfmixing laser Doppler velocimeter (SMLDV). In this paper, we theoretically obtained formulas giving lower and upper limit of measurable velocity. In the prototype digital vibrometer, the theoretical value was 6.7mm/s and 162.8mm/s, respectively, which agreed well with the measured value. The upper limit of measurable displacement amplitude was 12OO.mu.m at 10Hz, and 250.mu.m at 100Hz. Furthermore, the measurement accuracy the displacement amplitude was within 3% and average error 1.3%, when the shape of the sawtooth contained in the Doppler beat signal is clear and sharp. The measurement accuracy is found to depend on a degree of sawtooth asymmetry (DSA). 
Detecting a small threat object either fast moving or floating on shallow water presents a formidable challenge to shipboard sensor systems, which must determine whether or not to launch defensive weapons in a timely manner. An integrated multisensor concept is envisioned wherein the combined use of active and passive sensor is employed for the detection of short duration targets in dense ocean surface clutter to maximize detection range. The objective is to develop multisensor integration techniques that operate on detection data prior to track formation while simultaneously fusing contacts to tracks. In the system concept, detections from a low grazing angle search radar render designations to a sensorsearch infrared sensor for target classification which in turn designates an active electrooptical sensor for sector search and target verification.

In this paper, the fuzzy approximator and nonlinear inversion control scheme are considered. An adaptive nonlinear control is proposed based on the speed gradient algorithms proposed by Fradkov. This proposed control scheme is that three types of adaptive law is utilized to approximate the unknown function f by fuzzy logic system in designing the nonlinear inversion controller for the nonlinear system. In order to reduce the approximation errors, the differences of nonlinear function and fuzzy approximator, another three types of adaptive law is also introduced and the stability of proposed control scheme are proven with SG algorithm.

A method is presented for optimization of Universal Learning Networks (ULN), where, together with gradient method, Simulated Annealing (SA) is employed to elude local minima. The effectiveness of the method is shown by its application to control of a crane system.

A method is proposed which searches for optimal structures of Neural Networks (NN) using Genetic Algorithm (GA). The purpose of the method lies in not only finding an optimal NN structure but also leading us to the goal of selforganized control system that acquires its structure and its functionality by itself depending on its environment.

This paper evaluates the influences of sex on the human emotions while coexisting with robots. When we consider human vision, robot's motion is the most important parameter which influences human emotions and must be well controlled for males and females emotions. On the other hand, when we consider human touch of sense, which is effective for cooperation transmitting mutual forces, the softness of robot is an important parameter for human emotions and must be also well controlled for males and females emotions. From these points of view, at first, we evaluate robot's motion under four different shapes of velocity pattern while handing over a cup to humans. Second, we evaluate robot's softness realized by impedance control. From the first experiment, we concluded that the conditions of choosing an adequate maximum velocity value and locating the velocity peak at the center or the first half of the duration are necessary for male's emotions. In addition, the smooth velocity decrease in the last part of the velocity pattern's duration is desired for female's emotions. From the second experiment, we concluded that females prefer lighter values of virtual impedance characteristics than males and any small increase on the heaviness of virtual impedance values is followed by the negative exponential change on human emotions.

In this paper We propose one computer interface device for handicapped people. Input signals of the interface device are movements of eyeballs and head of handicapped. The movements of the eyeballs and head are detected by an image processing system. One feature of our system is that the operator is not obliged to wear any burdensome device like glasses and a helmet. The sensing performance of the image processing of the eyeballs and head is evaluated through experiments. Experimental results reveal the applicability of our system.

The authors experienced some industrial processes which essentially require human expertise. Human expertise for those processes have characteristics in common, and such human knowledge is described from a unified viewpoint in order to utilize it in automatization.

We investigate the applicability of the theory of robust stabilization with respect to additive, stable perturbations of a normalized leftcoprime factorization to controller design of a flexible arm with uncertain parameters.

In this paper, two representative schemes for vector control of induction motor without speed sensor are studied. First, the two sensorless systems which are implemented by voltage and current source are presented with new ideas and interpretations. Then a linear model around an operating point is proposed. Finally, the stability improvement of these systems are studied and evaluated by computing the trajectories of poles and zeros.

Furuhashi, Hideo;Shingu, Hiroyasu;Hayashi, Niichi;Watanabe, Shigeo;Sumi, Tetsuo;Uchida, Yoshiyuki 211
A precise openloop positioning system using linear pulse motor has been developed. The system is operated in a microstepping mode by controlling the electric current. One step of 508 .mu.m (tooth pitch of the linear pulse motor) is divided into 508 microsteps equally. The displacement is measured with a system using a Fiezeautype interferometer. Periodical positioning error with a period of the tooth pitch was observed in this system. Therefore, the position is corrected using the error. The error is stored into computer in advance, and the microstep current is corrected on basis of the stored data. Although the positioning error of the system without the correction was .+.4.5 .mu.m, that with the correction was decreased to .+.1.0 .mu.m. 
In this paper, a sliding mode controller (SMC) which can be characterized by high accuracy, fast response and robustness is applied to speed control of ACSERVO motor. The control input is changed to continuous one in the boundary layer to reduce the chattering phenomenon, and the boundary layer converges to zero when the state variables of system reach to steady state values. The integral compensator is added to reduce steady state error and to provide the continuous torque reference. The acceleration which is necessary to get the sliding plane is estimated by an observer. Sliding surface is included in control input to enhance the robustness and transient response without increasing sliding mode controller gain. The proposed controller is implemented by DSP(digital signal processor). The effectiveness of the proposed control scheme for speed controller is shown by the realtime experimental results in the paper.

This paper presents a control law of multiple actuation servo systems. Multiple actuation systems have an ability to solve some difficult engineering problems; Coulomb friction, backlash, and disturbance. This fact is shown by basic experiments as well as theoretical analysis. The proposed control strategy remarkably improves the performance comparing with conventional single actuation systems.

New output voltage control technique based on the simple feedback linearization is proposed. The system states are first divided into fast states and slow states. Then, the control stage is composed of the fast inner current control loop and the slow outer voltage control loop. From the inner loop, the average control is derived by the sliding mode concept and it is inserted into the dynamic equations of the slow states in the outer loop. Applying the feedback linearization technique to the obtained largesignal models of the PWM dcdc converters, linearized largesignal models are obtained for the slow states. With this technique, the output voltage controller of the PWM dcdc converters can be designed easily in the global state space and its control performance can also be much improved.

Stability of a coupled nonautonomous ordinary differential equation is investigated. Asymptotic convergence to zero of a part of state vector is additionally shown, otherwise only uniform stability could have been concluded by the Lyapunov direct method. Obtained results could be particularly useful in analysis of nonautonomous systems in which the invariance principle does not hold. An illustrating example is given.

In this paper, the frozen time approach is used to analyze the nonlinear system with time varying parameter. Using the extended linearization, we propose two analytical methods that compute an upper bound of the Euclidean norm of the difference between state variable and equilibrium point of the given system. The propertise of the two methods are discussed with simple examples.

It is useful to simulate the human visual function for the purpose of imageprocessing. In this study, the hardware of the spatial filter with the sensitivity of lateral inhibition is realized by the combination of optical parts with electronic circuits. The diffused film with the characteristics of Gaussian type is prepared as a spatial filter. An object's image is convoluted with the spatial filter. From the difference of the convoluted images, the zerocross position is detected at video rate. The edge of object is extracted in realtime by the use of this equipment. The resolution of edge changes with the value of the standard deviation of diffused film. In addition, it is possible to extract a directional edge selectively when the spatial filter with directional selectivity is used instead of Gaussian type of spatial filter.

In excavating tunnels, shield tunneling machines having many cutters on their cutter planes are used. Not many observation data being available in the detection system, optimal observation policy is very important. From this viewpoint, we previously considered the optimal location of acoustic sensors on the cutter plane and also the optimal observation policy for the case where three receiving transducers were used, and showed that the optimal sensor location was given as arbitrary equallyspaced points on the cutter plane circle, and that the optimal rotating angles were also found to be arbitrary. In application, however, it is often difficult to locate sensors at arbitrary positions or to use three sensors from the viewpoints of machine structure and cost. This paper considers the optimal observation policy for detecting anomlous plane objects for the case where two receiving transducers are used and the case where three receiving transducers are located only on a diameter of the cutter plane.

A technique is presented for evaluating spinal deformity of a human back by extracting a spinal line based on 3D topograpic reconstruction of the back from its moire image. A given moire image is differentiatedby DOG filter to extract moire stripes. The stripes are then assigned labels and the labels are interpolated by the Lagrange polynomial to yield the undulation of the back which gives a relative 3D shape of the back. A valley is searched on the undulation near the middle part of the back and the valley line is finally extracted as an approximated spinal line. The mean differenceand the variance between the spinal line and the middle line are calculated and reported. Experiment is performed employing real moire images ofjuniorhigh school students' backs and some of the results are shown with discussion.

Nowadays, the internal images of a human body can be easily provided by the ultrasound imaging, the Xray CT, or the MRI device, among which the ultrasound imaging device has good resolution for soft tissues of a human body compared with the other devices. Furthermore, the use of ultrasound imaging devices will increase in future especially in the obstetrics, territory, since it does not give harm to the human body. Although several techniques have been investigated until now in order to extract organs from ultrasound images, very few of them have achieved satisfactory results because of low contrast and high noise nature of images. This paper proposes a technique for automatic extraction of the gall bladder area from ultrasound images. The proposed technique first extracts a small reliable area of a gall bladder from an ultrasound image employing smoothing, binarization, expanding and shrinking, and labeling, and then expands the area referring to the binarized version of the original image. The technique is examined its performance by real ultrasound images of a gall bladder and satisfactory results are obtained. Some problems to be solved are discussed finally.

The purpose of this study is to provide the integrated process control system, utilizing neural network modeling, to search for the appropriate choice input, and to keep the process output within the desired rang in the real etch process.

In this paper, a model of an asymptotically reliable serial production line with quailty control devices is introduced and analyzed. By an asymptotic technique and Taylor series expansion, its average production rate is approximated in a closed form. The results are applied to a case study of a surface mount system.

Comtputer Aided Process Planning(CAPP) has been emerged as playing a key role in Computer Integrated Manufactunng(CIM) as the most critical link to integrate CAD and CAM. A modified variant CAPP system based on process planning rule base is developed in this paper. This CAPP system generates process plans automatically according to the GT code data provided as input. In order to execute process planning, various process planning rules are constructed in the form of decision tree and the inference engine that extracts the process plan based on the treestructured rules are implemented.

In this paper, a variable structure control scheme with a terminal sliding mode is proposed for robot manipulators. The proposed control scheme guarantees that the output tracking error converges to zero in finite time, and the overall system shows robust property against parametric uncertainties and external disturbances all the time.

This paper is concerned with the problem of robust stabilization of uncertain singleinput and singleoutput nonlinear systems. Based on the input/output linearization approach for nonlinear state feedback synthesis in conjunction with Lyapunov methods, a stabilizing state feedback controller is proposed. Compared with the controllers reported in the control literature, instead of uniform ultimate boudedness, the controller proposed in this paper can guarantee uniform asymptotic stability of nonlinear systems in the presence of uncertainties. The required information about uncertain dynamics in the system is only that the uncertainties are bounded in Euclidean norm by known functions of the system state.

In this paper, we consider a dynamic shape control problem with an example of controlling a flexible beam shape. Mathematical formulations are obtained by employing the Green's function approach. Necessary conditions for optimality are derived by considering the quadratic performance criteria. Numerical results for both of the dynamic and the static cases are obtained and compared.

In this paper, a moving horizon control algorithm, which can be applied for a wide class of nonlinear systems with control and state constraints, is considered. In a neighborhood of the origin, a linear feedback controller is applied. Outside this neighborhood, a moving horizon control law is applied. The time taken to solve an optimal control problem is considered in the algorithm so that the proposed control law can be applied as an online controller.

Due to finite bandwidth of missile dynamics, guidance commands in PN guidance tend to diverage as the missile approaches to the target. In this paper, a new method based on the shorttime stability theorem is introduced to extend the stability region.

Recently H
$_{\infty}$ control theory for nonlinear systems based on the HamiltonJacobi inequality has been developed. In this paper, we apply the state feedback controller solved via Riccati equation to a semiactive suspension model, two degree of freedom vehicle model, and show that it is effective for vibration control.. 
In this paper, two algorithms for computing multiple or clustered eigenvalues are proposed. The algorithm can be applied to all kinds of Hermitian matrix unlike the existing algorithm. Characteristics of the proposed algorithms is examined by MATLAB simulations.

In this paper, one simple technique to calibrate the system setting of the threedimensional measuring system is presented. Due to this technique, the threedimensional shape of the huge structures and the buildings can be readily obtained. This technique is applied to the threedimensional landscape simulation. Two examples are shown in this paper.

Recently, we proposed a 3D rangeimage measuring system for a slowly moving object by mechanically scanning a laser light beam emitted from a self mixing laser diode. In this paper, we introduced that every object moves along a straight line course, which is set diagonally against the semiconductor laser beam so that we can recognize each shape and size parameters of objects separately from the acquired 3D rangeimage. We measured a square mesa on a square plane as an object. The measured velocity was 4.44mm/s and 4.63mm/s with an error of 0.56mm/s to 0.37mm/s. And thickness error of the mesa was 0.5mm to 0.6mm, which was obtained from the 3D rangeimage of the standstill or moving object with thickness of 17.Omm.

The present paper deals with extraction of figures and characters from their background using the knowledge of color. At each pixel of the image on the CRT sent from a video camera, RGB values are transformed into the values in another color system, HSI, where "H" denotes hue;"S" denotes saturation;"I" denotes intensity. Representing color in HSI color space is advantageous, since a human feels color mainly in hue with the aid of brightness and purity. Comparing HSI data thus obtained with the masked original image detects noisefree edges included in the orginal image. Then setting a set of HSI thresholds and changing it identifies the portion of image of the same color. This color information is used in recongnizing characters and figures as an auxiliary system of a hierachical figure categorization method for characters and figures recognition.cters and figures recognition.

In the conventional systems, a human must have knowledge of machines and of their special language in communicating with machines. In one side, it is desirable for a human but in another side, it is true that achieving it is very elaborate and is also a significant cause of human error. To reduce this sort of human load, an intelligent manmachine interface is desirable to exist between a human operator and machines to be operated. In the ordinary human communication, not only linguistic information but also visual information is effective, compensating for each others defect. From this viewpoint, problem of translating verbal expressions to some visual image is discussed here in this paper. The location relation between any two objects in a visual scene is a key in translating verbal information to visual information, as is the case in Fig.l. The present translation system advances in knowledge with experience. It consists of Japanese Language processing, image processing, and Japanesescene translation functions.

This paper presents a method for measuring tool wear parameters based on two dimensional image information. The tool wear images were obtained from an ITV camera with magnifying and lighting devices, and were analyzed using image processing techniques such as thresholding, noise filtering and boundary tracing. Thresholding was used to transform the captured gray scale image into a binary image for rapid sequential image processing. The threshold level was determined using a novel technique in which the brightness histograms of two concentric windows containing the tool wear image were compared. The use of noise filtering and boundary tracing to reduce the measuring errors was explored. Performance tests of the measurement precision and processing speed revealed that the direct method was highly effective in intermittent tool wear monitoring.

The optimum design problem of a coathanger die is solved by the inverse formulation. The flow in the die is analyzed using threedimensional model. The new model for the manifold geometry is developed for the inverse formulation. The inverse problem for the optimum die geometry is formed as the optimization problem whose objective function is the linear combination of the square sum of pressure gradient deviation at die exit and the penalty function relating to the measure of nonsmoothness of solution. From the several iterative solutions of the optimization problem, the optimum solution can be obtained automatically while producing the uniform flow rate distribution at die exit.

As a trend toward multiproduct batch processes is increasing in Chemical Process Industry (CPI), multiproduct batch scheduling has been actively studied. But the optimal production scheduling problems for multiproduct batch processes are known as NPcomplete. Recently Ku and Karimi [5] have studied Simulated Annealing(SA) and Jung et al.[6] have developed Modified Simulated Annealing (MSA) method which was composed of two stage search algorithms for scheduling of batch processes with UIS and NIS. Jung et al.[9] also have studied the Common Intermediate Storage(CIS) policy which have accepted as a high efficient intermediate storage policy. It can be also applied to pipeless mobile intermediate storage pacilities. In spite of these above researches, there have been no contribution of scheduling of CIS policy for chemical batch processes. In this paper, we have developed another MSA for scheduling chemical batch processes with searching the suitable control parameters for CIS policy and have tested the this algorithm with randomly generated various scheduling problems. From these tests, MSA is outperformed to general SA for CIS batch process system.

A new air knife system for coating thickness control in hot dip galvanizing process had been developed and installed on the CGL in Pohang Steel Works, POSCO. This new system consists of air knives with remotely adjustable nozzle slot and an automatic control system which can control both longitudinal and traverse coating deviations. Based on the optimal control algorithm, a traverse coating deviation control was designed. The controller controls the lip profile of the air knives with flexible structure according to the deviation of coating weight. From the measured values which are dependent on the strip width, the lip gaps are calculated with optimal algorithm and the model of the coating deviation. Time delay between knives and a coating thickness gauge is solved by the Smith Predictor.

A successful, computeraided design support system can help a process designer focus on making effective design decisions, not merely tedious routine calculations. Such a system is essential to enhance quality of design in terms of economics, environmental benignity, reliability, robustness, and operability. Such a statement is even more accepted when applied to conceptual design problems, where gross design specifications are given while a combinatorial number of design alternatives exists. This paper presents an agentbased approach as a systematic and efficient way to design a design support system for the synthesis of conceptual chemical processes. An agentbased approach allows us to handle design knowledge as an object and thus greatly improve the modularity and reusability of that knowledge. Such modularity and reusability lead to the increased productivity in the development of a design support system and the increased ease in the relaxation of design decisions and the generation of design alternatives, both of which functions are critically important in dealing with the complexity and uncertainty of conceptual design problems.

A simple and effective method for improving Euclidean norm condition number for chemical processing system is presented. The singular value sensitivities of Freudenberg et al. (1982) is used to estimate the behavior of singular values of process transfer function matrix when design parameter is changed, then the condition number can be calculated straightforwardly. The method requires explicit dependencies of each transfer function matrix elements on design parameters. These dependencies can be obtained either by symbolic differentiation in the form of explicit function of design parameters, or by numerical perturbation studies for units with large and complicated models. Gerschgorintype lower bound for minimum singular value is introduced to detect the large divergencies near singular point due to linearity of sensitivities. The case studies are performed to show the efficiency of the proposed method.

An approach to efficient implementation of realtime control systems is presented in this paper. A compiler for translation of control algorithms is used in combination with a general program for realtime control. The compiler translates control algorithms written for the simulation in a design language to an implementation language. The translated algorithms are then automatically incorporated in the realtime control program.

In this paper, we get a reduced order controller in
$H^{\infty}$ mixed sensitivity problem with weighting functions. For this purpose, we define frequency weighted coprime factor of plant in$H^{\infty}$ mixed sensitivity problem and reduce the coprime factor using the frequency weighted balanced truncation technique. The we design the controller for plant with reduced order coprime factor using Jlossless coprime factorization technique. Using this approach, we can derive the robust stability condition and achieve good performance preservation in the closed loop system with reduced order controller. And it behaves well in both stable plant and unstable plant.t. 
The standard estimation and filtering theory are well known and has recently been incorporated with the H
$_{\infty}$ optimization techniques where the parametrizations of all estimators and filters are utilized. The issue of reducing its order is always of interest. This paper presents a method for synthesizing loworder stable state estimators. The method presented in this paper is based on the utilization of a free parameter function contained in the parametrization of all state estimators. The results obtained in the paper are compared with standard results on loworder estimators. Both results are shown to be the same in a sense of its orders, but the approaches taken are largely different. It is also shown in the paper that the method can easily and directly be extended to the Kalman filters and the H$_{\infty}$ (sub)optimal filters. Consequently, the orders of all state estimators, Kalman filters, and H$_{\infty}$ filters are shown to be reduced down to the number of states minus the number of outputs, respectively.ly. 
This note proposes a robust LQR method for systems with structured real parameter uncertainty based on Riccati equation approach. Emphasis is on the reduction of design conservatism in the sense of quadratic performance by utilizing the uncertainty structure. The class of uncertainty treated includes all the form of additive real parameter uncertainty, which has the multiple rank structure. To handle the structure of uncertainty, the scaling matrix with block diagonal structure is introduced. By changing the scaling matrix, all the possible set of uncertainty structures can be represented. Modified algebraic Riccati equation (MARE) is newly proposed to obtain a robust feedback control law, which makes the quadratic cost finite for an arbitrary scaling matrix. The remaining design freedom, that is, the scaling matrix is used for minimizing the upper bound of the quadratic cost for all possible set of uncertainties within the given bounds. A design example is shown to demonstrate the simplicity and the effectiveness of proposed method.

In this paper, estimation error bounds of the optimal FIR (Finite Impulse Response) filter, which is proposed by Kwon et al.[1, 2], are presented in discretetime systems with the model uncertainty. Performance bounds are here represented by the upper bounds on the difference of the estimation error covariances between the nominal and real values in case of the systems with the noise or model parameter uncertainty. The estimation error bounds of the discretetime optimal FIR filter is compared with those of the Kalman filter via a numerical example applied to the simulation problem by Toda and Patel[3]. Simulation results show that the former has robuster performance than the latter.

This paper uses a new method to improve the performance criterion of an active suspension car. The used control strategy is based on robust H
$_{\infty}$ control theory taking into consideration the chasis flexibility. It will be shown that the modeling errors can be lumped into an unstructured uncertainty and the robust controller designed in the presence of these perturbations could maintain the stability and performance even for the controlled true system.. 
In this paper, we describe the composition of frequency response bands based on experimental data of plants (controlled systems) with uncertainty and nonlinearity, and the robust stability evaluation of feedback control systems. Analysis and design of control systems using the upper and lower bounds of such experimental data would be effective as a practicable method which is not heavily dependent upon mathematical models such as the transfer function. First, we present a method to composite gain characteristic bands of frequency response of cascade connected plants with uncertainty and a recurrent inequality for the composition. Next, evaluation methods of the robust stability of multiloop control systems obtained through feedback from the output terminals and multiloop control systems obtained through feedback into the input terminals are described. In actual control systems, experimental data of frequency responses often depends on the amplitude of input. Therefore, we present the evaluation method of the nominal value and the width of the frequency response band in such a case, and finally give numerical examples based on virtual experimental data.

In this paper, a design method of learning flight control system via input matching is proposed. The proposed learning control system is a simple structure which has an artificial neural network and feedback mechanism, and it is a useful method to control nonlinear systems.

This paper describes dynamic analysis and attitude control of a large spacecraft with flexible appendages in gravitational field. The effect of attitude control and vibration control of flexible appendages in gravitational field has been clarified. We demonstrate some simulations in gravitational field for some cases, and suggest the effects of gravitational torque, parameters of flexible appendages, attitude control and vibration control of flexible appendages.

The LQG/LTR controller design procedure for ground alignment of inertial platform is accomplished. Due to the alignment system dynamics, LQG/LTR controller is proposed to overcome both singular problem and nonsquare problem. To show the effectiveness of this control system, computer simulation was performed under the assumption of random sway motion.

To force an aircraft to track the specified path, the generation of the smooth desired trajectory is essential. In this paper, the cubic spline function is used to generate the trajectory which passes through the specified intercept points. The simulation results show that the desired trajectory generated by the spline interpolation is very smooth and the aircraft tracks it with small position errors.

Since the generation and transmission of telecommand in satellite monitoring and control system depend on the decisions of operators, it is possible that operators with different levels of knowledge may generate different telecommands in the same situation. Because of this reason, automation technology of satellite operation is being researched and developed to minimize the decision error due to the operator's lack of experience. This paper suggests a method of automated satellite control, which generates telecommands automatically using the knowledge of satellite subsystem engineers or specialists for the ground system. This method provides safe satellite operation and expansion of satellite life time by automatic generation of the telecommands, so that the operator's interrupt is minimized which provides the efficient satellite control.

A helicopter is used in a variety of situations because of its usability. Its operation, needs human skill. The authors are working on automatization of human skill. Helicopter operation is one of such fields of practicing human skill. This is why the present paper deals with helicopter (model helicopter) operation. Full operation of a helicopter needs more complicated system in both aspects of software and hardware, and also requires more training for operation. From the purpose here that helicopter operation is for checking the applicability of the authors' idea for automatization based on experience, attitude regulation in hovering is the target. In the present paper, a human operator's operation is recorded as a time series of operation actions, and the record is reorganized as the correspondence between the helicopter's attitude and the proper operation action required in that particular situation.

This paper describes computer control of a wheel chair by using landmarks. Firstly, the approach of landmark detection and recognition is described and the image coordinates are obtained by the primary component analysis method. Subsequently, the selflocalization of the wheel chair is determined on the basis of a threedimensional image processing method. Finally, the control system of the wheel chair is described and a navigation experiment is given. Experimental results indicate the effectiveness of our approah.

This paper presents a time efficient method for determining a sequence of locations of a mobile manipulator that facilitates tracking of continuous path in cluttered environment. Given the task trajectory in the form of octree data structure, the algorithm performs characterization of task space and subsequent multistage optimization process to determine task feasible locations of the robot. Firstly, the collision free portion of the trajectory is determined and classified according to uniqueness domains of the inverse kinematics solutions. Then by implementing the extent of task feasible subspace into an optimization criteria, a multistage optimization problem is formulated to determines the task feasible locations of the mobile manipulator. The effectiveness of the proposed method is shown through a simulation study performed for a 3d.o.f. manipulator with generic kinematic structure.

This paper presents a new formulation of the kinematics of closedchain mechanisms and its applications to obtaining the kinematic solutions and analyzing the singularities. Closedchain mechanisms under consideration may have the redundancy in the number of joints. A closedchain mechanism can be treated as the parallel connection of two openchains with respect to a point of interest. The kinematics of a closedchain mechanism is then obtained by imposing the kinematic constraints of the closedchain on the kinematics of the two openchains. First, we formulate the kinematics of a closedchain mechanism using the kinematic constraint between the controllable active joints and the rest of joints, instead of the kinematic constraint between the two openchains. The kinematic formulation presented in this paper is valid for closedchain mechanisms with and without the redundancy. Next, based on the derived kinematics of a closedchain mechanism, we provide the kinematic solutions which are more physically meaningful and less sensitive to numerical instability, and also suggest an effective way to analyze the singularities. Finally, the computational cost associated with the kinematic formulation is analyzed.

A robot system is proposed to realize coordinated motion of two arm robot. Due to a 3D vision sensor, precise coordinated motions could be realized. Using a sophisticated IC chip, real time image processing could be executed using a simple circuit.

Teleoperating system has been developed for several decades, and many control schemes for it have been suggested. But the implementation for real application needs very simple but effective controller. In this paper, an advanced control scheme for this purpose is suggested, which is the combination of a modified internal model controller and variable filter for force reflection. And we verify the effectiveness of the proposed scheme through the experiment. We use PUMA560 as the slave robot, which is operated by velocity servo loop with geared motor. Both the responses of free motion and contact motion are shown.

In this paper, we discuss the force control of flexible manipulators. Since the force control of flexible manipulators with planar one or two links using the distributedparameter modeling has been the subject of a considerable number of publications until now, real time computations of the force control schemes are possible. But, application of those control schemes to multilink spatial manipulators is fairly complicated. In this paper, we apply a concise hybrid position/force control scheme for a flexible manipulators. We use a lumpedparameter modeling for the flexible manipulators. The Hamilton's principle is applied to derive the equations of motion for the system and then, statespace model is obtained by the Lagrange's method. Finally, comparison of simulation results with experimental results is given to show the performance of our method.

Previously obtained results of L
$_{2}$ gain and H$_{\infty}$ control via state feedback of nonlinear systems are extended to a class of nonlinear system with uncertainties. The required information about the uncertainties is that the uncertainties are bounded in Euclidian norm by known functions of the system state. The conditions are characterized in terms of the corresponding HamiltonJacobi equations or inequalities (HJEI). An algorithm for finding an approximate local solution of HamiltonJacobi equation is given. This results and algorithm are illustrated on a numerical example.. 
A finitedimensional approximation technique is developed for a class of spectral systems with input and output operators which are unbounded. A corresponding bounding technique on the frequencyresponse error is also established for control system design. Our goal is to construct an uncertainty model including a nominal plant and its error bounds so that the results from robust linear control theory can be applied to guarantee a closed loop control performance. We demonstrate by numerical example that these techniques are applicable, with a modest computational burden, to a wide class of distributed parameter system plants.

The stability of system is one of the important aspects and to judge system's stability is another complicated problem. Previously, new technique derived from relaxing Lyapunov conditions has been already introduced and in this paper, this proposed technique applies to the practical dynamic systems. This utility of numerical procedures prove the comparable improvements of the estimation of robustness for dynamic systems having structured (bounded) perturbations.

In this paper, a control law based on the receding horizon concept which robustly stabilizes timevarying discrete linear systems, is proposed. A dynamic game problem minimizing the worst case performance, is adopted as an optimization problem which should be resolved at every current time. The objective of the proposed control law is to guarantee the closed loop stability and the infinite horizon
$H^{\infty}$ norm bound. It is shown that the objective can be achieved by selecting the proper terminal weighting matrices which satisfy the inequality conditions proposed in this paper. An example is included to illustrate the results.. 
In this paper we propose an online tuning method by using genetic algorithm for robust minimax IPD controller based on new criterion. The new criterion is the Integral of Squared Error (ISE) with a penalty of the derivative of manipulated variable. The work focuses on robust tuning of IPD controller's parameters in the presence of plant parameter uncertainty. The result of several simulation studies are provided to illustrate the performance of this robust tunig method.

In this paper a new approach to obtain the solution of the linearquadratic Gaussian control problem for singularly perturbed discretetime stochastic systems is proposed. The alogorithm proposed is based on exploring the previous results that the exact solution of the global discrete algebraic Riccati equations is found in terms of the reducedorder pureslow and purefast nonsymmetric continuoustime algebraic Riccati equations and, in addition, the optimal global Kalman filter is decomposed into pureslow and purefast local optimal filters both driven by the system measurements and the system optimal control input. It is shown that the optimal linearquadratic Gaussian control problem for singularly perturbed linear discrete systems takes the complete decomposition and parallelism between pureslow and purefast filters and controllers.

In this paper we present a method to construct fuzzy model with multidimension input membership function, which can construct fuzzy inference system on one node of the network directly. This method comes from a common framework called Universal Learning Network (ULN). The fuzzy model under the framework of ULN is called Universal Learning Networkbased Fuzzy Inference System (ULNFIS), which possesses certain advantages over other networks such as neural network. We also introduce how to imitate a real system with ULN and a control scheme using ULNFIS.

A control strategy for random parametic system by using FLC is investigated and developed. In the course of research, as a first part, nonlinear system under random disturbance is investigated. Preliminary results is presented in the paper. A control technique, GA based FLC, is employed successfully for inverted pendulum experiencing white noise excitation.

The satellite in a circular orbit about a planet with disturbances and a threedegreeoffreedom (3DOF) structure under seismic excitations are modeled by the linear stochastic differential equations. Then the risksensitive optimal control method is applied to those equations. The mean and the variance of the cost function varies with respect to the risksensitivity parameter, .gamma.
$_{RS}$ . For a particular risksensitivity parameter value, risksensitive control reduces to LQG control. Furthermore, the derivation of the mean square value of the state and control action are given for a finitehorizon fullstatefeedback risksensitive control system. The risksensitive controller outperforms a classical LQG controller in the mean square sense of the state and the control action. 
A general task execution with hybrid impedance control method is addressed. The target impedance is expressed in the constraint frame. For the computational simplicity and the robustness improvement, disturbance observer scheme is used. To make stable contact with the environment, the large value of desired inertia gain for the forcecontrolled subspace is suggested. Numerical examples are given to show the performance of the proposed controller.

Various physical limitations which intrinsically exist in the manipulator control system, for example kinematic limits and torque limit, cause some undesirable effects. Specifically, when one or more actuators are saturated the expected control performance can not be anticipated and in some cases it induces instability of the system. The effect of torque limit, especially for redundant manipulators, is studied in this article, and an analytic method to reconstruct the control input using the redundancy is proposed based on the kinematically decomposed modeling of redundant manipulators. It results to no degradation of the output motion closedloop dynamics at the cost of the least degradation of the null motion closedloop dynamics. Numerical simulations help to verify the advantages of the proposed scheme.

The problem of compliant motion control using a redundant manipulator is addressed in this article. Specifically, a hybridcontrol type and impedancecontrol type controllers are extended to general redundant manipulators based on the kinematically decomposed and geometrically compatible modeling of its joint space. In the case of the hybrid controller, it leads to the linear and decoupled closedloop dynamics in the three motion spaces, that is the motioncontrolled, forcecontrolled, and the null motioncontrolled spaces of the redundant manipulator. When the proposed impedance controller is applied, the decoupled impedance models in three motion spaces are obtained. The superiority of the proposed controllers is verified with the numerical experiments.

The present paper studies a robot manipulator's contact tasks on the uncertain flexible objects. The flexible object's distributed parameter model is approximated into a lumped "position statevarying" model. By using the wellknown nonlinear feedback compensation, the robot's control space is decomposed into the position control subspace and the object's torque control subspace. The optimal state feedback is designed for the position loop, and the robot's contact force is controlled through controlling the resultant torque on the object using modelreference simple adaptive control. Experiments of a PUMA robot interacting with an aluminum plate show the effectiveness of this control approach. approach.

Autonomous Mobile Robot(AMR) is a field of study which is under active research along with rapid development of the engineering technology. The main reasons for the high interest in AMR are because of its ability to change work space freely and its capability to replace human being for difficult and dangerous jobs. Also the fact that AMR provides a variety of research fields, such as path planning, navigation algorithm, sensor fusion, image processing, and controller design is part of the reason for its popularity. But relatively few researches are concerned with controller. So in this paper, a control strategy of mobile robot with nonholonomic constraint for tracking ordered discontinuous motion is proposed. The proposed control strategy has been designed as a state feedback shape to allow the AMR to obtain continuous velocity and track the path which is composed of discontinuous motions. In order to design such controller, 3 states have been reduced to 2 states through coordinate projection. These ideas are tested for validity through simulation and simulation result is compared with experiments result.

In this paper, we propose a new strategy for a space robot to control its attitude. A space robot is an example of a class of nonholonomic systems, a system of which cannot be stabilized into its equilibria with continuous static state feedbacks even in the case that the system is, in some sense, controllable. Thus, we cannot design stabilizing controllers for space robots using conventional control theories. The strategy presented here transforms the nonholonomic system into a timestate control form, and allows us to make the state of the original system any desired one. In the stabilization, any conventional control theory can be applied. For simplicity, a space robot with a twolink manipulator is considered, and a simulated motion of the controlled system is shown.

In this paper, we investigate socalled the falling cat problem. It is well known that a cat, when released from an upside down configuration starting from rest, is able to land on her feet without violating angular momentum conservation. This has being an interesting problem for engineers for a long time. We consider a model of a falling cat as connected two rigid columns, which is a nonholonomic system. We design the controller for it, using time state control form of the model and exact linearization technique. Finally, we test the controller thorough simulation on the model of a falling cat.

A real time and adaptive method for obtaining Volterra kernels of a nonlinear system by use of pseudorandom Msequences and correlation technique is proposed. The Volterra kernels are calculated real time and the obtained Volterra kernels becomes more accurate as time goes on. The simulation results show the effectiveness of this method for identifying timevarying nonlinear system.

Identification of a process in closedloop control system is an important problem in practice. This paper deals with parameter estimation using inputoutput data of the process operating in a closedloop system. It is necessary to determine orders and delaytime to get consistent estimators by least square method for inputoutput data collected from the process. The authors considered a problem to determine delaytime in the condition that orders were known, in last KACC. So we extend the range to determine orders and delaytime in this paper.

An interacting multiple model (IMM) approach which merges two hypotheses for the situations of constant speed and constant acceleration model is considered for the tracking of maneuvering target. The inflexibility of uncertainty which lies in the kinematic constraint (KC) represented by pseudomeasurement noise variance is compensated by the mixing of estimates from two model Kalman tracker: one with KC and one without KC. The numerically simulated tracking performance is compared for the "great circular like turning" trajectory maneuver by the single model tracker with constant speed KC and two model tracker which is developed in this paper.his paper.

The paper presents a new global maximum search method for multimodal unknown functions of two variables. The search method is composed of two stages and sequentially samples the candidate point in a subdomain selected using a priority function in each stage. The search domain is autosimilarly divided into triangular subdomains, or cells, during the search process. A measure of accuracy of local maximum search is introduced to check if a local search has converged to a specified accuracy or the maximum of a local peak cannot be the global maximum. A criterion for switching from the first to the second stage, is proposed using a ratio of the observed peak width to the largest cell in the domain. By numerical simulations, the required number of trials is evaluated for some function models with different peak parameters, and the switching criterion is optimally determined. The results show that the proposed method obtains global maximum points with certainty and saves largely computation time even for functions with extremely steep peaks.

A simple method to design observers for linear describtor systems with unmeasureable disturbance is represented by the response of a linear free system. The sufficient conditions for the existence of the observer are given. The design procedures of an identify and a minimal order observers are shown, respectively.

In this paper, a new detection scheme, the detectable maneuver set (DMS) scheme, is proposed by incorporating the tradeoff property between target maneuver magnitude and detection time delay. With this new detection scheme, small maneuvers can be effectively detected without enlarging window size. Simulation results show that the proposed DMS scheme gives better tracking performance.

A neural network based upon the back propagation algorithm was designed and applied to acoustic power spectra of electrohydraulic total artificial hearts in order to diagnose mechanical failure of devices. The trained network distinguished spectra of the mechanically damaged device from those of the undamaged device with overall success rate of 63%. Moreover, the network correctly classified more than 70% of spectra in the frequency bands of 0100 Hz and 700950 Hz. Consequently, the neural network analysis was useful for the diagnosis of mechanical failure of a total artificial heart.

The automatic interpretation of awake background electroencephalogram (EEG), consisting of quantitative EEG interpretation and EEG report making, has been developed by the authors based on EEG data visually inspected by an electroencephalographer (EEGer). The present study was focused on the adaptability of the automatic EEG interpretation which was accomplished by the constructive neural network with forgetting factor. The artificial neural network (ANN) was constructed so as to give the integrative decision of the EEG by using the input signals of the intermediate judgment of 13 items of the EEG. The feature of the ANN was that it adapted to any EEGer who gave visual inspection for the training data. The developed method was evaluated based on the EEG data of 57 patients. The retrained ANN adapted to another EEGer appropriately.

A microcontrollerbased DC motor control system for a total artificial heart(TAH) was developed. Using a onechip microcontroller, 87Cl96KB, the design of digital motor speed control system and servo control system is demonstrated. Functionally, the control system consists of a position control unit, a speed control unit, and a communication unit. The performance and the reliability of the developed control system were assessed through a series of mock circulation system experiments.

A new stroke output control algorithm with a fuzzy logic for an electrohydraulic left ventricular assist device(EHLVAD) was developed. The EHLVAD pumps out blood from left atrium actively. Excessive suction of blood may cause fatal damage in left atrium. The LVAD has to provide a maximal stroke output without collapse of left atrium. In this study a new fuzzy algorithm for predicting and detecting suction and doing proper action on LVAD without using an extra pressure sensor but with bellows pressure signal and motor current signal is developed. The performance of the fuzzy control algorithm is demonstrated by the results from mock circulatory experiments.

In this paper, a learning control problem is formulated for cooperating multiplerobot manipulators with uncertain system parameters. The commonly held object is also assumed to be unknown and the multiplerobots themselfs experience uncertain operating conditions such as link parameters, viscous friction parameters, suctions, actuator bias, and etc. Under these conditions, the learning controllers designed for learning of uncertain parameters and robot control inputs for multiplerobot systems are shown to drive the multiplerobot manipulators to follow the desired Cartesian trajectory with the desired internal forces to the unknown object.

A manipulation of a multiple contacted object by a Rotational Base and Singlejointed Finger mechanism(RBSF mechanism) is discussed. The manipulation is characterized by multiple contacts on an object and large motions of the object with sliding contacts. The kinematics and dynamics allowing sliding at multiple contacts are explored. The conditions for manipulation of an object at multiple contacts by the RBSF mechanism, which cannot exert arbitrary contact forces because it has a fewer number of joints than is required for active control, is presented.

The closed loop robots have the advantage of higher velocity capability and often higher precision in comparison to the open loop robots. We have simulated the kinematic analysis of three limbs robot (TL robot), which is one of the closed loop robots. After then, we have designed experimental TL robots with 3 different kind of actuators. In this paper, we described the experimental results, and the problems in its applicatons.

This paper deals with the problem of motion planning for a unicyclelike robot. We present a simple local planner for unicycle model, based on an approximation of the desired configuration generated by local holonomic planner that ignores motion constraints. To guarantee a collision avoidance, we propose an inequality constraint, based on the motion analysis with the constant control input and time interval. Consequently, we formulate our problem as the constrained optimization problem and a feedback scheme based on local sensor information is established by simply solving this problem. Through simulations, we confirm the validity and effectiveness of our algorithm.

In this paper, we propose a control strategy for a class of nonholonomic systems. A system with nonholonomic constraint is called a nonholonomic system, and as Brockett showed, the equilibrium of such systems can not be stabilized with any continuous static state feedbacks even though the system is controllable in the sense of nonlinear. A control strategy we propose is transforming this system into timestate control form by coordinate transformation and input transformation. We will apply this control strategy to the motion control of a rigid ball that is held between two parallel plates.

In steel works, reheating furnace is an essential part of a rod mill plant and it treats various types of billets continuously. Although getting an optimal setting for a single billet is simple, control setting for whole groups of billets is a difficult task. In this work, we studied a detail mathematical model and optimal control setting of reheating furnace. As the mathematical model of each billet is a partial differential equation, online control is almost impossible for the whole billets charged into the furnace. Therefore, we tried to provide a guideline for optimal setting value of the roof(index) temperature for the target billets which account for about 20% of the charged billets.