Omnidirectional visual information is important to ensure the safety for the autonomous navigation of mobile robots. Usually such vision systems use a fish-eye lens or a reflecting mirror (spherical or hyperboloidal). The former system has a problem that unnecessary visual field occupies the most part of image. The latter one has also a problem that the camera image is blurred due to the physical property of the mirror curvature and thus the clear image cannot be obtained especially in a dark environment. In order to solve the above problems, we propose omnidirectional vision system using two reflecting mirrors with contrived curvature to obtain a panoramic image. Theoretical consideration is made to estimate the degree of aberration for the reflecting optical system that uses two mirrors with arbitrary curvatures, and a design method is developed to minimize the aberration of the image. The optical characteristics of a design example are evaluated and the results show the effectiveness of the proposed vision system.

View full abstract

This paper newly proposes a telescopic microphone system that may establish the high sensitivity and SNR, and narrow directivity in 3D-space, by using passive dynamic focusing and spherical wave synthesis with a spherical sensor array. On this system, passive dynamic focusing and spherical wave synthesis are realized by compensating the amplitude attenuation of detected signals, due to wave propagation, with variable gain elements, adjusting the relative time delays among the signals with variable delay elements, so as to match the resultant signals with those propagated hypothetically from the source at a predetermined focused point, and summing up the resultant signals. As the result, the sensitivity of the system is enhanced, for a sound source at the focused point, by a multiplicative factor of the number of used microphones. For uncorrelated additive detection noises, the SNR of the system output is improved, for the source at the focused point, approximately by the same factor as the sensitivity. After the sensitivity 3D-distribution of the system and the SNR characteristics of a synthesized output being made clear, the relative RMS error spatial distribution of the system output is examined through computer simulations by assuming a FM wide-band source signal which covers audio frequency bands. All the results of theoretical or numerical analysis and computer simulations illustrate the effectiveness as well as the fundamental characteristics of the proposed telescopic microphone system, resulting in the possibility of constructing the intended telescopic microphone system.

View full abstract

The second order approximate model (i.e., equivalent vibrating piston model) is obtained for the circular pressure detecting elements such as an edge-clamped membrane and a similar thin plate. The spring constant of the model is defined by the force for making the unit average-displacement of the detecting element. The mass of the model is not the real value of the element, but the effective mass modified by the first order eigenvalue of the element. For a membrane the effective mass is given by 8m^m/α_m1²(≈1.3833m_m) where m^m is the real mass and α^ml is the eigenvalue. For a plate the effective mass is 192m_p/α_p1⁴(≈1.8400m_p) where m_p is the real mass and α_p1 is the eigenvalue.
A dynamical model for pressure sensor systems, which is composed of the radiation impedance of the vibrating piston in a infinite baffle plate contacted with gas or liquid, the second order model mentioned above and the impedance of an air cavity behind the vibrating piston, is also presented in order to analyze the static and dynamic characteristics of the system. The natural frequencies, the static displacement gain and the static pressure gain of the pressure sensor systems are obtained by the use of the dynamical model.

View full abstract

It is difficult to measure the flow rate of compressible fluids especially the unsteady flow, because the density of them changes largely with the temperature. Therefore, there is no effective method for testing the dynamic characterisitics of air flowmeters. If there is an apparatus which could generate the unsteady flow, the dynamic characterisitics of air flowmeters could be measured. It is known from our previous study that the isothermal condition during air discharge or charge can almost be realized by stuffing the steel wool into the chamber. Then, the flow charged into or discharged from the chamber becomes only the function of pressure. Hence, the both steady and unsteady flow rate is obtained by differentiating the pressure in the chamber during air charge or discharge.
In this study, an apparatus which could generate the desired flow rate of compressible fluids using the isothermal chamber is proposed. The flow rate can be generated by controlling the pressure in the chamber during air discharge with a servo valve. Both the steady and unsteady oscillating flow could be generated with the apparatus. The performance of the generator was examined by experiments and confirmed its effectiveness.

View full abstract

In this paper we presented a method of high precision load measurement by two freedom dynamic crane scale
The basic point of this approach lies in the idea that the sampling averages of dynamic load signal during a time interval between the maximum swing angles and other central parts of the time interval are functions of the maximum swing angle and the minimum swing angle. The mass of goods is a constant factor of these functions. By the sampling averages, the mass value of goods can be calculated under the dynamic condition. The simulation results indicated that the measurement precision of more than 0.1% F.S. can be achieved.

View full abstract

A new method for signal transform by use of M-sequence, called M-transform, is proposed, and some properties of the M-transform are described. The essence of M-transform is to suppose any time signal to be the output of a filter whose input is an M-sequence. The application of the M-transform to linear system identification is described, and the result of computer simulation show a good agreement with the theoretical consideration.

View full abstract

Inside wall of blast furnace is frequently defected by its high pressure and temperature process and solid bodies like iron ore and cokes. For the safety and efficient operations of blast fuanaces, state-based detection of the defect of the inside wall is of great importance. One of the authors previously proposed a detection system which used an ultra-sonic sensor and measured the thicknesses of the entire and inside walls.
The paper presents an improved system which measures the wall-thicknesses even when many reflected waves from the inside of the wall are complicatedly superposed under a time-variant unknown bias caused by the measurement process. The system also makes use of the phase infomation of the reflected waves to raise the reliability of the measurement result. Finally, the effectiveness of the proposed system is verified with experiments on a furnace-type steeve and an actual blast furnace.

View full abstract

Particle Imaging Velocimetry (PIV), which is the whole flow field measurement technique using visualized images, has been recognized to be essential and very useful for analyzing two-or three-dimensional complex flow fields. Although lots of measurement methods for PIV have been reported, the PIV is clasified to the pattern tracking and the particle tracking methods in principle. The pattern tracking method such as the correlation method and the gray level difference method gives some erroneous vectors because of mismatching a tracer image pattern to another one. While velocity vector profiles measured by the particle tracking method such as the four-step PTV (Particle Tracking Velocimetry), the spring-model method, and the binary image correlation method, partly lack information on velocity vectors because no tracer particles exist in some parts of the flow field. It is required, therefore, to estimate a correct velocity vector field from the measured vectors with erroneous ones or to estimate velocity vectors in the unmeasurable parts. One promising approach for the estimation is to make an appropriate model of the field by using the measured information.
This paper proposes a new method using artificial neural network for estimating the whole flow velocity vector fields from measured velocity vectors. The neural network is trained by using measured velocity vectors as teaching data so that the derivatives of a certain scalar function agree well with the measured data. The continuity equation of flow is consequently satisfied in the estimated vector fields and the scalar function gives the stream function. The merit of this method is that measured data on velocity vectors are automatically corrected and the estimated data satisfy the continuity equation of flow. The pattern of streamlines is additionally obtained.
The method is applied to the PIV standard images developed based on a calculated velocity field that is available through the Internet (http://sap.gen.u-tokyo.ac.jp) in order to evaluate the effectiveness and acuracy. It is proved that the method gives much correcter velocity vector distributions than the measured ones.

View full abstract

In this paper, the robust tracking performance for a class of uncertain plants with triangular structure is considered. A decentralized feedback configuration is adopted in order to preserve this triangular structure in the closed loop system and enable the asymptotic tracking with as less servomechanism as possible. The necessary and sufficient condition for the robust asymptotic tracking subject to this decentralized configuration is derived. Synthesis based on the obtained condition will yield servomechanism of order lower than that obtained by the Internal-Model Principle.

View full abstract

The problem of adaptive robust stabilization for a class of uncertain single-input and single-output (SISO) nonlinear systems is considered. It is assumed for the uncertainty to be bounded. The bound, however, may be unknown. Based on Lyapunov method and input/output linearization approach, adaptive control laws are developed so that no prior knowledge of the bounds on the uncertainties is required. By updating these upper bounds, we desgin a class of continuous state feedback controllers. It is shown that under the proposed controller with adaptive laws the states of the controlled system converge to zero as time approaches infinity. Because of its continuity, our continuous controller can guarantee the existence of the solution to the dynamical system in usual sense. Finally, a numerical example is presented to demonstrate the effectiveness of the approach developed in this paper.

View full abstract

This paper is concerned with a synthesis of an integral type robust control system based on a non-standard H_∞ control problem: it is an H_∞ control problem for a generalized plant where the direct feed-through term of a subsystem from external inputs to measurement outputs is column full rank. There are many approaches to designing the integral type robust control system, however they are somewhat complex. The purpose of this paper is proposing a simple approach based on the formulation to the non-standard H_∞ control. Integral type robust controllers for plants: integral type and non-integral type, can be designed with the proposed approach.

View full abstract

Recently, the input-output linearization technique of affine nonlinear systems has been attracted a lot of researchers. Especially for multivariable systems, the nonlinear version of the structure algorithm plays a basic role. In linear systems, the structure algorithm is the equivalent concept to the interactor. However, since the interactor contains more information about the delay structure of multivariable systems than the structure algorithm, the interactor is often used instead of the structure algorithm in many of linear multivariable control systems. The purpose of this paper is to define an interactor for affine nonlinear multivariable systems and to use it for the input-output linearization. It will be shown that this interactor is invariant under the nonlinear state feedback and many properties of the interactor for linear systems can be naturally extended to the nonlinear case.

View full abstract

This paper describes a graphical evaluation of the robust stability in a frequency domain based on the results from our previous papers in which Popov's criterion was expressed in an explicit form. The control system described herein is a feedback system with one time-invariant nonlinear element (a sector nonlinearity) in the forward path. Considering the application to a computer-aided control system design, we will present an evaluation method of the robust stability in connection with the size of sector nonlinearity and the gain margin on a gain-phase diagram (i.e., a modified Nichols chart). We will show two results as numerical examples: one is where Aizerman's conjecture was approved, and the other is where it was not.

View full abstract

The objective of this paper is to propose an optimal control strategy for a linear time-invariant plant in the input-output data space. The control strategy does not use any traditional mathematical model such as a transfer function or a state-space equation. The plant dynamics is represented as a set of basis vectors whose elements are input-output data of the plant. Using this representation, two optimal control problems are solved. One is to find the control input which minimizes a quadratic performance index. The other is to find the control input which minimizes a quadratic performance index subject to achieving dead-beat tracking.

View full abstract

The main objective of this paper is to present a necessary and sufficient condition for stability of linear time-invariant descriptor systems in terms of a linear matrix inequality (LMI) with no equality constraint. The stability condition is used to consider robust stabilization of systems with polytopic uncertainties via linear descriptor variable feedback. To apply the LMI approach to systems with uncertainties in the coefficient matrix for the derivative of the descriptor variable, the system is augmented introducing a new variable so that no uncertainty appears in that coefficient matrix and the robust stability property is preserved. A gain of stabilizing feedback is computed using the solution of the LMI for robust stabilization.

View full abstract

In this study, we propose a class of dynamic compensators based on linear function observers which achieve arbitrary pole assignment of the closed loop. The construction procedure directly utilizes Wang's algorithm for the static output feedback pole assignment problem, and the same technique can be applied to dual observers also. As a result, assuming the structure of either observer, the sufficient condition of the compensator's order is naturally derived. The condition is more transparent than the one proposed by Rosenthal and Wang, and it is shown that both give the same minimal order under certain assumptions.

View full abstract

LSDP (Loop Shaping Design Procedure) is one of the most promising design procedures in H_∞ control theory. It has many special features except a problem that its central controller tends to have higher order. In this paper, we investigate some possibility on the existence of lower order controllers in the set of LSDP controllers. If our proposed conditions are satisfied, the order of controller is same as the total order of weighting transfer matricies. Confirming the validity of our result, we apply it to the robust vibration control of a flexible beam with an electromagnet as actuator and gyro sensor which can detects angular velocity of beam. In this case, our proposed conditions are satisfied and a second order controller is obtained. As the controller is in the set of LSDP controller, it has the almost same characteristics as the central controller which has 10th order. Using the Balanced truncation method, the lower limit is 6th order.

View full abstract

The new type of the design method proposed in this paper is similar to the one using phase lead-lag compensator and has two particular characteristics. First, the approximate system described real order lag element with a dead time element is specially adopted to identify the controlled object. The principle of the identification derived from the frequency characteristics is presented. Secondly, the compensator is composed by an integral element and phase lead-lag elements having multiple zeros and multiple poles. The order of phase lead-lag elements is, in general, an integer but in this design it will be treated as a real number. In order to use practically the compensator, the impulse response function of the compensator is obtained in the form of a series expansion of a gamma function and it can be performed by the convolution method between the impulse response function of the compensator and a deviation error signal.
The compensator parameters are determined by a rule of trial and error in order to obtain a quick response using manipulating variable under the maximum value assigned in advance. Since the order of the compensator can be adjusted finely through numerical treatment as to a real, the quick response by a relatively small manipulating variable has been performed. Then, the selection of the sampling time in the case of practical digital control is considered in detail. The effectiveness of this control design method is shown by examples.

View full abstract

Diophantine polynomial matrix equation is essential tool in the analysis and synthesis of the linear control systems. In general, the inverse of a large dimensional real matrix is necessary to solve the equation. In this paper, a solution of Diophantine equation will be considered using an extended division algorithm. This method needs the inversion of n-th dimensional real matrix, where n is the system degree, and this number is smaller than the ordinally methods. Using the solution, we will obtain another description of all stabilizing controllers.

View full abstract

In this paper, we discuss a simple distributed parameter controller for large space structures with lumped and distributed flexibility. As a simple example of large space structures, we consider two flexible beams connected by a spring. Dynamic equations of the distributed parameter system are derived by means of the Hamilton's principle. We introduce the Lyapunov function related to the total energy of the distributed parameter system and derive a simple sensor output feedback control law. Using the characteristic of the differential operator and the invariance principle, we prove the asymptotic stability of the closed-loop system and the convergence property to the desired stationary state. In order to demonstrate the validity of the derived model and the proposed controller, experiments have been carried out.

View full abstract

The Electrical Discharge Machine(EDM) which can process metals of any hardness, is generally used for making moulds and dies. During the process, because the process is non-contact, unlike milling, drilling or grinding, a large machining force does not seem to exist. However, when using a large electrode, a large machining force is generated by the viscosity of the dielectric liquid. Such force deforms the machine and makes the machined depth several tens to hundreds of micrometers less than the required depth. As a result, additional machining for the shallow areas is required, which obviously increases total machining time.
To cope with this problem, we have proposed a new method to utilize a disturbance observer to compensate for the discrepancy in the depth, according to the estimated value of the deformation. The disturbance observer is designed based on a mechanical model identified by modal analysis, in which a machining force is considered as step disturbance. With the proposed method in an actual machining, although the value of deformation is up to about two hundreds micrometers, the error in the depth is kept within a few micrometers. It does not need any additional machining for realizing the required depth, and the total machining time is about 40% shorter than for the conventional method. This result confirms that the proposed method is essential for accurate and fast machining where a large machining force is present.

View full abstract

Automatic Landing FLight EXperiment, ALFLEX, is conducted by National Aerospace Laboratory (NAL) and National Space Development Agency of Japan (NASDA) in order to develop the automatic landing technology for future unmanned reentry space vehicle, HOPE (H-2 Orbiting PlanE). Total of 13 flight experiments were conducted at Woomera Australia in 1996, and all flight experiments were perfectly completed.
Because ALFLEX vehicle, which is dynamically similar sub-scaled model of HOPE, is the unmanned and autonomous vehicle which has tip-fin wings, the flight control law has to achieve high-response-performance and robustness simultaneously. In this paper, the design method, which connects the MDM/MDP (Multiple Delay Model/Multiple Design Point) and the H_∞ control design method combined with Exact Model Matching to achieve the objectives, is presented. And the results of flight experiment about the performance and the robustness are also presented.

View full abstract

In this paper, we propose VLSI architecture of Kalman Filters with minimum latency based on FADDEEVA algorithm. It is important to minimize latency, because latency degrades a stability of feedback systems. In proposed architecture, the minimization of latency is achieved by parallel computation of parameters of Kalman Filters and the triangularization and nullification process of the FADDEEVA algorithm. This architecture has a highly parallel structure. For an n state and m measurement dynamic system, this architecture requires latency of only (3+m+6⌈log₂n⌉)t_as+(6+m)t_m+t_d, where t_as, t_m and t_d indicate delay of an adder, a multiplier and a divider, respectively. This means a reduction of latency of approximately 0.06 times when compared to a systolic array Kalman filters for the filter order of m=n=100. Moreover, latency of this architecture slightly increases against the increasing filter order, so that this architecture is also effective to the higher order Kalman filters. The proposed architecture is very suitable for VLSI implementation, because it can be simply constructed from six kinds of modules and some basic arithmetic elements.

View full abstract

A partial language allows us to describe the behavior of Petri nets correctly. In a partial language, refinement means replacing an event by some partial language. In this paper, we first define a hierarchical partial language which ensures the same refined partial language without the distinction of order of replacement. We then study a condition/event net construction problem with specification described by a hierarchical partial language and present a solition technique.

View full abstract

The conventional optical flow gives apparent motion in the two dimensional image plane. An estimation of 3D velocity field including occluded parts without mixing tracer etc. to the fluid had not only never been proposed but also impossible by the conventional computer vision algorithm. In this paper, we propose a new method of three dimensional optical flow of the fluid based on physical model, where some boundary conditions are given from a priori knowledge of the flow configuration. Optical flow is obtained by minimizing the mean square errors of a basic constraint and the matching error terms with visual data using Euler equations. Here, Navier-Stokes motion equations and the differences between occluded data and observable data are employed as the basic constraints. First, we verify the effectiveness of our proposed method by applying our algorithm to simulated data with partly artificially deleted and recovering the lacking data. Next, applying our method to the fluid of observable surface data and the knowledge of boundary conditions, we demonstrate that the 3-D optical flows are obtained by the proposed algorithm.

View full abstract

This paper describes a new approach to the X-ray computed tomography in order to find defects of the products in manufacturing industry. Our approach which is Fast Model Reconstruction Computed Tomography (FMR-CT), involves two key ideas: 1) for forward problem, integration two dimensional sinc functions along the X-ray on the sampled two-dimentional planee model and 2) for inverse problem, solving the liner inverse problem using a truncated singular value decomposition (SVD) method.
Our algorithm is favorable fast computation capability because image vector can be obtained simple formula such as pseudo-inverse matrix (constant) multiplied by projection data vector, and is also compatible with multiprocessor implementations. The combination of sampling theorem and the SVD allows smooth reconstruction image with even smaller numbers of projections, because each techniques contribute interpolation and least-squares optimization, respectively.
We demonstrate the cross section image which is reconstructed real projection data sets acquired by using the micro-focused X-ray TV. Further, we discuss practical determination of truncation rank with the visualization of decomposed v-singular vector and the distribution of singular values.

View full abstract

Although a lot of human errors has occurred in the maintenance of power plants, systematic approaches to the reduction of human error are not sufficient compared with those applied to operations. The authors propose a new method of evaluating and analyzing human interface design from the viewpoint of human error reduction, and have implemented the method on the DIAS system. This system consists of maintenance personnel and equipment interface simulators, and it generates a dynamic interaction between humans and the working environment. The DIAS system can calculate the human error rate and their weighted calculations according to dangerousness, the influence on the equipment, and the influence on the plant of each task by THERP, and can carry out the dynamic analysis of the panel selection error index if there is similar equipment in the neighborhood, the working time, the distance of movement, and the distance of eye movement. The authors applied this system to evaluate the interface design of actual Transformer Protection Relay Panels and their layout in a room in a nuclear power plant. Our customer accepted our evaluation and proposals modifying the panel design.

View full abstract

This work is to propose a concept of intelligent human-machine interface for process control that can provide a proper display of information based on intention inferencing of the plant operator. Operator's intention inferred in the goal hierarchy of operational tasks is used to select a proper display format for the specific information channel of the display. The concept was applied to a DURESS (Dual Reservoir System Simulation) plant and an experiment for evaluation of the interface design was conducted. The experimental result showed that the automatic selection of display format by intention inferencing outperformed manual selection in operator performance and the intelligent interface proposed is effective for operation aid.

View full abstract

The initial state of controllers is discussed in terms of improving the transient. We derive an optimal value of the initial state, which attains favorable transient of the closed loop system.

View full abstract

This paper considers decentralized stabilization of large-scale systems composed of a number of interconnected subsystems, the information structure constraint of which is compatible with the subsystems. The decentralized stabilization problem is reduced to a feasibility problem of a bilinear matrix inequality (BMI). To solve the BMI, an algorithm is proposed using the idea of the homotopy method, where the interconnections between subsystems are increased gradually from zeros to the given magnitudes.

View full abstract