The chronology of books on science and technology, particularly mechanical engineering, which significantly contributed to the successful transfer of technology from the Western countries to Japan around the time of Meiji Restoration, is traced and the extent to which each book contributed to the modernization of Japan is assessed. A genealogical tree is used to classify the books according to their origin, level of quality, and where and how they were used.

抄録全体を表示

A new method for simple robust control of multimodes of vibration of flexible structures is proposed. Although robust control theories such as H_∞ theory are useful for controlling systems with unknown factors and parameter variations, they are not effective for controlling multimodes of vibration of flexible structures. The proposed method combines the suboptimum control theory and two kinds of filtering technique for reducing the number of sensors and preventing spillover of neglected higher modes of vibration. Therefore, the method is called the filtered suboptimal control method. The effectiveness of the method is demonstrated by showing that the first five modes of vibration of a model structure of large bridge tower are well controlled without spillover of neglected higher modes.

抄録全体を表示

This paper is the 3rd of a 4 report series which proposes an active method for controlling the vibration of main towers of huge bridges due to strong wind. In the first and second report, the modeling procedure and the control strategy based on suboptimal control theory and filter technique are described and verified by simulation and impulse experiments. In this study control experiments in an artificial wind environment are carried out: a scaled bridge tower model is placed in a wind tunnel and excited by turbulent wind flow with speed varying from 0 m/s to 13 m/s. It is confirmed that the first five modes of bending and twisting vibrations are well controlled by the proposed method.

抄録全体を表示

The vibration of towers of suspended bridges due to strong wind has been of great concern to scientists and engineers. In this study, attention is given to the vibration control of a tall bridge tower under construction. An active control method is presented based on the lumped model of the bridge tower and the suboptimal control theory. Since the dynamic parameters of the tower differ under different phases of construction, it is necessary to reevaluate the corresponding model, nevertheless, the control strategy is maintained. Three typical construction phases of a bridge tower are considered in this paper. The modeling method and results of the simulation control are presented to illustrate the validity of the proposed method.

抄録全体を表示

This paper describes the suppression of vibration which is generated on rotating machinery with an overhung rotor. We propose a rotor-shaft system supported by rubber springs which are installed on the pole surfaces of the electromagnets. This is a hybrid type equipment in which a bearing housing is supported elastically by rubber springs and controlled actively by electromagnetic actuators. These actuators are controlled by displacement sensors at the pedestal and/or the roller and can apply an electromagnetic force which suppresses any vibration of the roller. In the analysis, we calculate the controller gains which minimized the amplitude of vibration at the roller. In the experiment, the results correspond to the analytical results, and we confirm that the hybrid type equipment is a highly effective method for vibration suppression.

抄録全体を表示

A new method of self-sensing magnetic bearings is introduced. Magnetic bearings are intended to function both as a forcing actuator and a gap sensor. The magnet has two coils. One is a bias coil which provides constant bias flux to the air gaps. The other is a control coil which increases one gap flux and degreases the other gap flux. The magnetic force is linearized to the control coil current using this technique. The coils are also used as a differential transformer-type displacement sensor. The current of the bias coil is adjusted by changing the dusy ratio of the pulse-width modulated signal. The induced carrier frequency component in the control coil is detected and put into the demodulating circuit to obtain the rotor displacement. Two sets of driving and demodulating circuits are used to detect the horizontal and vertical displacements of the rotor. The position of the rotor is controlled using the proposed method. The rotor can run up to 12000 rpm using the self-sensing magnetic bearings.

抄録全体を表示

This study is concerned with the adaptive forced rejection approach for a single periodic output disturbance with an unknown period and its application to unbalance vibration suppression in rotor-magnetic bearing systems. A discrete recursive technique is applied to estimate the unknown disturbance period and then to internally and adaptively generate a pseudofeedforward reference signal such that this disturbance is eliminated. Simulations and demonstration experiments are carried out. The results indicate that both the estimation algorithm of an unknown period and reference signal adaptive generation algorithm are stable. Finally, the approach is applied to a problem of unbalance vibration suppression in rotor magnetic bearing systems. Approximately 20dB and 7dB reductions in unbalance vibrations were experimentally obtained in the horizontal and vertical directions of the front and rare sides of the rotor, respectively. The experimental results show that the proposed algorithm is effective for achieving unbalance vibration suppression.

抄録全体を表示

In this paper we describe the coupled mode superposition method for the magnetic damping problem, which is produced by the coupled effect between the deformation and the induced eddy current of the structures for future fusion reactors and magnetically levitated vehicles. The formulation of the coupled mode superposition method is based on the matrix equation for the eddy current and the structure using the coupled mode vectors. Symmetric form of the coupled matrix equation is obtained. Coupled problems of a thin plate are solved to verify the formulation and the computer code. These problems are solved efficiently by this method using only a few coupled modes. Consideration of the coupled mode vectors shows that the coupled effects are included completely in each coupled mode.

抄録全体を表示

The purpose of this study is to analyze the nonlinear vibration phenomena of vibration cutting systems. Vibration cutting often suffers from problems caused by the nonlinear vibration phenomena; therefore it is necessary to analyze these phenomena by applying a nonlinear dynamic model. We propose a new vibration cutting model described by nonlinear differential equations, which is based on Grabec's cutting model. Numerical study of the model indicates that the model successfully replicates the characteristics of actual vibration cutting systems: the pulsating cutting force and the reduction of the amplitude of the vibration of work.

抄録全体を表示

A new boring bar whose stability against chatter is superior to not only conventional steel bars but also cemented carbide bars, has been successfully developed. A high-modulus carbon-fiber reinforced plastic was used as the material of the bar. Four types of bar having different shapes of steel core were tested while carbon fibers were aligned unidirectionally in the longitudinal direction to give high bending stiffness for all cases. The bar with a cross-shaped steel core shows the best cutting capability and stability among all CFRP boring bars. It can be used when the ratio between length (L) and diameter (D) is 7 even under severe conditions, whereas a cemented carbide bar sometimes chatters even at L/D=6. The cross-shaped steel core can increase the bending stiffness of the bar in both tangential and radial directions by constraining the shear deformation of fiber layers without sacrificing resonant frequencies significantly.

抄録全体を表示

The full-closed-loop is popularly used to improve the positioning accuracy of machine tools and electrical discharge machines. However, it consists of a non collocated motor and position sensor, and thus the elastic deformation between the motor and the sensor induces the mechanical vibration which restricts the bandwidth of positioning control. To suppress this instability, we proposed a new controller which feeds back the relative displacement and the relative velocity from the motor to the positioning table. A stability condition is investigated in the 2-DOF model that is the basis of the full closed-loop system, and the precise effect of the controller is examined in a 4 DOF model of an actual electrical discharge machine. Operation of an actual machining test clarified that machining speed was enhanced twofold in the finishing machining in the lateral servo mode, and the positioning accuracy was maintained to be equivalent to that of a usual full-closed-loop.

抄録全体を表示

The calculation time of constrained multirigid-body dynamics by conventional methods is proportional to the cube of the model size. Compared with this, the calculation time of Order-N-Algorithm is directly proportional to the model size. Rosenthal showed this method, but his paper explained only step by step procedure of an example, and thus it was hard to grasp the whole aspect of the algorithm. The mathematical formulation of the algorithm, and simple examples of investigations about the calculation load are given in this paper.

抄録全体を表示

In order to automate the scooping, it is desirable that the resistive force acting on the bucket is obtained before the scooping is performed. The objective of this paper is to derive theoretically the resistive force acting on the bucket during scooping of piled rock fragments. First, the changes in shape of the rock pile due to the penetration of it by the bucket are modeled. Based on this model, the resistive force acting on the bucket is considered theoretically. Finally, it was confirmed that the calculated results obtained using the model agree well with the experimental ones.

抄録全体を表示

Signal processing of multifrequency channel decompositions should be possible in the near future because signal processing for each frequency channel is easy. However, a difficult problem, i.e., aliasing, exists when the signal is reconstructed. The wavelet transform technique has a special property, i.e., that the signal can be reconstructed uniquely after multifrequency channel decompositions. In this study, a new practical identification method, in which the wavelet transform technique is used, is proposed. This proposed method is used to identify a one degree of freedom pneumatic robot which is developed in this study. The applicability of the proposed method is confirmed by the experimental results.

抄録全体を表示

We report the measurement of shock pressure which cannot be measured directly using a normal pressure transducer. We considered three types of shock absorber: steel, brass and Teflon, and we measured their transfer ratios using a shock tube. Results show that Teflon is the most effective shock absorber because the transfer ratio is very small. Our procedure is as follows: (1) unknown pressure is measured through the shock absorber (damped pressure); (2) the spectrum is calculated by fast Fourier transform (FFT); (3) the real pressure spectrum is reproduced using the transfer ratio; (4) the real pressure wave is reproduced by inverse FFT. We found that the reproduced pressure agrees well with the directly measured pressure. And as an example of a pressure which cannot be measured using a pressure transducer we measured the center pressure of a convergent detonation. From the measurement we estimate the pressure to be about 360 MPa.

抄録全体を表示

Using the discharge-type blade tip clearance measurement system which was designed for testing, the following were clarified. According to Paschen's law, gas pressure exerts an effect on the measured values; however, the result of the experiment showed that the measured values were not affected by pressure with the minimum discharge voltage. It was confirmed that when the speed of the probe was high, sometimes the probe came in contact with the blade tip. Since there exists a time constant for voltage restoration in discharge signals, there is a limit beyond which the continuous signals can not be obtained. This was examined theoretically and confirmed experimentally. As a result of measuring the blade tip clearance at a peripheral speed of 400 m/s, the elongation of a rotating object due to peripheral speed could be measured.

抄録全体を表示

In order to investigate problems involving soil and construction machinery such as a hydraulic excavator, understanding of the physical interactions between the soil and the machine is essential. The purpose of this study is to experimentally examine actual excavation processes by examining various physical parameters, for the purpose of realizing an intelligent excavation machine whose performance is not only automatic but also optimal in terms of energy consumption, operation time and excavation trajectory. In this article we present experimental results using a model-sized bucket and two different soil types, sand and loam. Various sensing devices and experimental apparatus were developed for the experiments. Excavation load and stress distributions were examined.

抄録全体を表示

The handling force of a car in motion is usually measured by a device for measuring torque acted on steering column spindle. However, this measurement method is influenced by the inertia force of the steering wheel. In some cases, the inertia force is accompanied by an extremely large discrepancy. Therefore, we designed a method for measuring the handling force which actually acted on the steering wheel surroundings and which was uninfiuenced by the inertia force of the steering wheel. In this paper, we describe this measurement method and the results of measurement using an instrument made for experimental purposes.

抄録全体を表示

Humans have excellent visual cognition ability, which allows them to recognize dot patterns and their clustering, but the mechanism behind this ability has not yet been clarified. When a human observes particles made up of several subgroups, he has a different sense of aggregation depending on the state of the particles. This paper proposes a new technique for the recognition and evaluation of particle aggregation. We subdivide the influencing regions of particles using mask operations. Thus, a state in which particles are distributed is represented with a density distribution chart. Based on this chart, we defined three measurements: dense area, total density and dense intensity. Several experiments are conducted in order to test the performance of the developed measurements. It is verified that the dense intensity approximates human recognition of particle aggregation well.

抄録全体を表示

A new ambulance stretcher that decreases blood pressure variation (BPV) of a patient is discussed. BPV is caused by foot-to-head acceleration in the patient's body during an emergency brake, and may cause serious damages to a patient suffering from circulatory disease. We propose a method to suppress foot-to-head acceleration in patient using the gravitational acceleration of the ambulance. The tilt angie of the stretcher is continuously controlled according to the back to front acceleration of the ambulance when an emergency brake is applied. The back-to-front acceleration is measured by an acceleration transducer located on the floor of the ambulance. A computer derives the desired angle of the stretcher from the back-to-front acceleration. Emergency brake tests were performed using a fixed stretcher and the stretcher proposed here. The results show that the foot-to-head acceleration is canceled by tilting the stretcher, and BPV obtained using the proposed stretcher is less than half that obtained using the fixed stretcher.

抄録全体を表示

This paper deals with human body vibration while gazing at a moving picture. It is said that one gets seventy percent of the information from the external world through one's eyes. We consider that this percentage increases especially when one is gazing at a moving object. One acts reflexively according to the influence of the information received. If one makes a mistake while gazing at moving surroundings under seismic or vibrating floor conditions, a severe accident may occur. This is related to human operating error in an emergency. In this paper, we investigate how one acts reflexively when one is gazing at a moving picture. For a person standing on a force plate gazing at a vibrating picture generated by a personal computer in a dark room, we measured the position of his center of gravity. As a result, it was found that the resonant frequency of a human being is between 0.1 and 0.3 Hz.

抄録全体を表示

Data input devices for physically handicapped people have already been developed. One simple device detects the blowing action of the operator as input signals to select desired menu items. Another advanced device detects eyeball movements as input signals. One problem associated with these advanced data input devices is that the operator is often obliged to wear a sensor or device. In this work a data input device operated using eye movements is presented. An important feature of our system is that the operator is not obliged to wear a sensing device.

抄録全体を表示

An ultrasonic method for flaw type determination is described. Clearly, the lower the frequency of an ultrasonic longitudinal wave becomes, the wider the beam spreads. In the case of a large flaw, there is a difference in the beam path between low and high frequency components. Therefore, the propagation time of a low frequency component is longer than that of a high frequency component. On the other hand, in the case of a small flaw, there is little difference in the propagation time between the components. Hence it is possible to determine the flaw type, by analyzing the propagation time of a particular frequency range. Time-frequency, t-f, analysis is applied to the ultrasonic signals reflected by three types of artificial flaws; side drilled holes, fiat-bottomed holes and slits. The Short Time Fourier Transform, STFT, method of t-f analysis is found to be effective. The results support the feasibility of classification of flaws by means of t-f analysis of the reflection of ultrasonic waves emitted by probes of different diameter.

抄録全体を表示

This paper considers the modal filtering of a planar structure using point sensors designed for extracting a particular modal amplitude of the structure. It is purpose of this paper to establish the design procedure of a modal filtering based upon point sensors. Much as a distributed modal sensor outstrips a point sensor in the sensing potential, the handling and maintenance are odious. To cope with this problem, this paper presents a design methodology of modal filtering using point sensors easy to handle and maintain it. It is found that modal filtering design of point sensors can be achieved by simply following the design procedure of distributed modal sensors already presented by the authors. It is also found that point sensors perfectly replace the distributed sensors for modal filtering in the frequency range of interest. Experiment is conducted, demonstrating the validity of the point sensor-based modal filtering.

抄録全体を表示

As technology improves, the spacing between magnetic disks and magnetic heads continues to be reduced. A typical spacing measurement method is based on interferometry, in which the magnetic disk is usually replaced by a glass disk. Therefore, this method cannot be used to measure the spacing between a real magnetic disk and a ceramic slider. In order to measure this real spacing, we have developed a new spacing sensor based on "Frustrated Total Reflection". We fabricated the sensor on a ceramic slider using the magnetic head insulation layer as a waveguide. To evaluate the performance of the new sensor, we used it to measure the spacing between a glass disk and a ceramic slider, using a typical interferometry method. The results agreed with each other within an error of 5 nm. The sensor shows potential for use with a real magnetic slider.

抄録全体を表示

An application technique for micro drops with diameter as small as 100 μm is desired for repairing the micro patterns of electronic circuits. In this paper we propose a micro drop application system that utilizes visual feedback control. A micro object handling robot, which is equipped with optical microscopes and a "task package", transfers micro drops one by one from the reservoir to the desired position on a substrate by means of a needle shaped tool, and applies them while controlling their diameter based on real time visual measurement. The experiment proved that the diameter of micro drops of metal paste can be controlled within an accuracy of 4%.

抄録全体を表示

In the automation of arc welding using visual sensors, it is necessary to precisely detect the weld line. However, for large scale structures, such as the steel tube transmission towers examined in the present study, it is not easy to maintain the reliability of weld line detection. Therefore, assuming that welding data among structural parts, including positions and directions of their contact points and lines, can be easily obtained from CAD data, and utilizing the fact that the current torch position is known all the time, we have proposed image processing algorithms to reliably detect only the welding area with a specified direction, and to precisely detect the welding line for torch position control. The proposed algorithms have been applied to a test piece placed on an experimental system, and continuous tracking along the welding line by measuring the welding gap has been presented.

抄録全体を表示

This paper deals with the driving torque property of a rib-mesh antenna for satellite use. Experiments are used to verify the proposed prediction method. The antenna is driven by an antenna deployment mechanism (ADM) with a motor. Driving torque measurement is performed at both ADM and antenna levels. Measured results show that the driving torque increases at a dramatic rate at angles between 80 and 83.4 degrees then decreases at angles over 83.4 degrees. Discussion given to clarify torque degradation shows that the prediction method is effective.

抄録全体を表示

A method of determining the operational strategy for energy supply systems in consideration of equipment startup/shutdown costs is proposed. This operational planning problem is formulated as a large-scale mixed-integer linear programming problem, in which the on/off status and energy flow rates of equipment are determined so as to minimize the sum of operational and startup/shutdown costs over the period considered. By utilizing a special feature of the problem, an algorithm for solving the problem efficiently is proposed, in which dynamic programming is adopted to determine the most cost-effective path regarding on/off status, the branch and bound method is modified to enumerate multiple candidates for on/off status, and a strategy is incorporated to reduce the number of candidates by using lower and upper bounds for the objective function. Through a numerical study on the operational planning of an energy supply system for district heating and cooling, the effectiveness of the proposed algorithm is ascertained in terms of computation time, and the effect of startup/shutdown costs on the operational strategy and cost is investigated.

抄録全体を表示

Enclosures of machines have openings that allow radiation of heat and sound leakage. It has been reported that a sound insulating device is effective for dissipating sound. In this paper, in addition to the conventional method, we suggest a method of oscillating the sound insulating device using a piezoelectric actuator to control it actively. First, eigenvalue analysis of sound field within an enclosure is conducted using BEM. Then, the state equation for sound field is derived. Finally, it is proven by numerical simulation that this proposed active control system can control sound within an enclosure successfully.

抄録全体を表示

The flexible microactuator (FMA) is a new pneumatic rubber actuator which has been applied to various kinds of miniature robots. While conventional FMAs are reinforced with fiber in the circumferential direction to resist deformation in the radial direction, the fabrication of FMAs with homogeneous rubber and without fiberreinforcement is necessary in order to manufacture them at reasonable cost. In this work, an optimum design of an FMA without fiberreinforcement is accomplished using a nonlinear FEM based sensitivity analysis, so that the maximum stress and the radial expansion are decreased, and the bending angle is increased. The resultant deformation of the optimized FMA agrees well with the experimental results.

抄録全体を表示

This paper considers the modal excitation of a planar structure using point actuators designed for exciting a particular modal amplitude of the structure. It is the purpose of this paper to establish the design procedure of a modal excitation methodology based upon point actuators. Much as a distributed modal actuator outstrips a point actuator in modal excitation, a distributed actuator with sufficient power is non existent. To cope with this problem, this paper presents a design methodology of modal excitation using point actuators that are easy to obtain and handle. It is found that modal excitation using point actuators can be achieved by simply following the design procedure of distributed modal sensors already presented by the authors. It is also found that point actuators can perfectly replace the distributed actuators for modal excitation in the frequency range of interest. Experiment is conducted, demonstrating the validity of the point actuator-based modal excitation.

抄録全体を表示

For an underwater robot controller, nonlinearity in the system is a critical problem. The nonlinearity is too strong for control techniques based on a linear model even when robust controllers are designed. A sliding mode control technique is of considerable interest due to its capability to deal with nonlinear systems. However, there is no simple, suitable observer for the nonlinear model. In this study, focussing on a second power nonlinearity, a nonlinear model observer is derived by improving the variable structure systems (VSS) observer for a linear model. Using this observer, a positioning controller of an underwater flexible rotating arm driven by a thruster is developed to which the sliding mode control technique is applied. Through computer simulations, it is shown that the nonlinear model observer is able to accurately estimate the thrust of the thruster, which could not be estimated by the VSS observer. In addition, the performance of the developed control system is confirmed by positioning experiments.

抄録全体を表示

The technical application of a master slave system is a very attractive for micromanipulation and a power amplifier. In the master-slave system, the human being must instruct the slave by operating the master, in order to execute the tasks. In this paper, variable impedance control of the master arm is discussed on the basis of H_∞ control theory. Impedance control is useful for adjusting the dynamic characteristics of the master arm, such as the apparent decrease of inertia. However, it is difficult to determine the impedance parameters uniquely, since the desired impedance varies according to the operating conditions. Therefore, in view of maneuverability and stability, the damping factor of the reference impedance is varied according to information on acceleration and velocity. The aim of this study is to design a time varying controller dependent on the damping factor, based on gain scheduling. The gain scheduled problem can be solved via elementary manipulations on linear matrix inequalities (LMI). The validity of variable impedance control, in which an operation force is actively utilized, is confirmed in control experiments using a direct-drive master system.

抄録全体を表示

In this paper, we present a method for calculating the relative position and orientation of landmarks from a mobile robot using a camera. The method is based upon a triangular method for plural point features in space. To speed up its calculation time, we apply a formula that has closed form solutions using 3 point correspondences. To reduce the errors of the relative position, we introduce a fusion method using multiple points with optimal weighting matrices derived by considering minimum error variance. We also propose an algorithm for finding a common solution. The proposed method reduces the measurement errors and has the advantages of high speed measurement. To evaluate the proposed method, the error and calculation time are compared to those of several similar methods. The proposed method has the greatest advantage in the real-time measurement of mobile robots.

抄録全体を表示

We propose a strategy for cooperation between a position controlled manipulator and a crane with three wires. A compliant mechanism consisting of three free joints is designed and installed at the tip of the manipulator to avoid excessive inner forces due to mutual positioning errors of the manipulator and the crane. Since the cooperative system with the compliant mechanism becomes kinematically redundant, the characteristics of the redundancy are analyzed and a way to exploit the redundancy is discussed. The robot is position-controlled as a master and the crane is velocity-controlled as a follower by the feedback signals from the sensor systems of the crane and the free joint mechanism. The experimental results obtained using a prototype cooperation system verify the effectiveness of the cooperative strategy and the system architecture.

抄録全体を表示

Because of its good performance in suppressing parameter variations, its simplicity, and its usefulness for real-time control, the disturbance observer has recently been employed for manipulator control. However, how stably the disturbance observer suppresses parameter variations in the total manipulator control system has not been examined. In this paper we analyze a manipulator control system based on the disturbance observer. From the results of analysis, we propose a design method for the manipulator control system based on the disturbance observer. Finally, we present experimental results obtained from an implementation of the proposed method.

抄録全体を表示

In this study, the optimum arm posture of a coupled tendon-driven multijoint manipulator arm (or CT Arm) at maximum payload output was derived and the corresponding tendon traction forces were also analyzed, during management of a heavy payload by the manipulator in a gravity environment. The CT Arm is special tendon traction transmission mechanism in which a pair of tendons used to drive a joint is puIled from base actuators via pulleys mounted on the base-side joints. This mechanism enables optimal utilization of the coupled drive function of tendon traction forces and thus enables the lightweight manipulator to exhibit large payload capability. The properties of the CT Arm mechanism are elucidated by the proposed optimal posture control scheme. Computer simulation was also executed to verify the validity of the proposed control scheme.

抄録全体を表示

Robots are required to have the ability to adapt their function according to the tasks to be carried out in an unexpected environment, and to execute tasks even if a part of the system is malfunctions. Fault Tolerance is a significant factor of functional adaptability. In this paper, a fault tolerant control method with a proxy control strategy for a distributed manipulator is proposed. A Byzantine fault model is assumed in the method, where in the behavior of the faulty part cannot be predicted. The method focuses on malfunction of CPU (central processing unit) which is the controller of the manipulator. The method consists of procedures for fault detection, localization, containment, system reconfiguration and error recovery. The fault detection procedure is based on communication using shared memory. A voting algorithm for fault location is proposed. The fault tolerance control method is implemented in a distributed manipulator with modular architecture, called Fun-ARM (functionally adaptive and robust manipulator). A reaching motion experiment with a CPU pseudo fault is shown, and the proposed fault-tolerant control method is verified.

抄録全体を表示

Biological systems in living organisms can be regarded as the ultimate among distributed information processing systems and have been expected to provide feasible ideas to engineering fields, especially robotics. Among these systems, the immune system plays a role in protecting living organisms against countless foreign substances in order to cope with a dynamically changing environment. This ability is exactly what we need for an autonomous mobile robot. Therefore, in this paper, we focus on the immune system and apply it to a gait control problem of a 6-legged walking robot. We aim at realizing an action selection mechanism through a decentralized learning method inspired from the immune network architecture. The validity of our proposed method is confirmed by simulations.

抄録全体を表示

We present a new approach to genetic programming (GP) for robot learning. Our goal is to realize robustness in robot navigation control by means of the automatic discovery of functions. In robot learning by conventional GP, the program is not separated for each partial behavior because the program is represented as a long program. This results in lack of reusability and robustness in the program. Due to the fact that GP can not rearrange the acquired program if the environment is unstable. We propose an approach, called COAST, to improve the robustness, which has a library for storing certain subroutines for reuse. These subroutines are generated by automatically defined functions. We use the wall following problem to demonstrate the efficiency of this mehod.

抄録全体を表示

A home robot (a robot used in the home) must recognize the intention of a person in the home, in order to support that person. It is important to construct intelligence with which a robot can predict the person's next act and help him or her if required. In this paper, a robust system which predicts vague human actions is proposed. Human actions can be modeled by a stochastic process in digital time and discrete states: this stochastic process is visualized using a Petri net. A series of human actions is a chain of states and acts. A human state is expressed as a place, and a human act as a transition. A stochastic act is described as a predicting transition, and transition probability to plural states is given. If the system makes a wrong prediction, the transition probability is renewed. As time progresses, the level of this intelligence increases, and the prediction will be accurate with high probalility.

抄録全体を表示

The present paper deals with precise measurements of the diffusion coefficients as well as the saturated solubilities of single component gases such as N₂, O₂ and CO₂ to a mineral oil. The method to determine the diffusivity is based on measuring the pressure changes caused by the one-dimensional diffusion between the gas and the oil enclosed in an airtight container. For N₂ and O₂ the profiles of the measured pressure changes agree well with those predicted by the diffusion theory, whereas that is not the case with CO₂. Although the reason why CO₂ does not seem to obey the diffusion theory has yet to be studied, it may suggest the possibility that the diffusion coefficient varies with the pressure, considering that the range of pressure changes in the diffusivity measurement was much larger for CO₂ than for the other two gases. The scattering of the diffusion coefficients obtained for N₂ and O₂ by this method fell to within +30% of the average. Moreover the solubility measurements have revealed that Henry's law holds true between the three pure gases and the oil tested, and that O₂ and CO₂ dissolve into the oil approximately two and ten times more, respectively, than N₂.

抄録全体を表示

This paper presents a new modified Reynolds equation for ultrathin gas film lubrication in the whole range of Knudsen numbers. Weber's technique, which combines the viscous slip-flow equation and free molecule flow equation, is improved by adding another coefficient to describe more precisely Poiseuille flow rate between two parallel plates at an arbitrary Knudsen number. It is shown that the theoretical flow rate by the improved technique agrees well with Dong's experimental results in the wide range of Knudsen numbers from 0.003 to 900. The modified Reynolds equation using the theoretical flow rate is as simple in form and easy to handle in numerical analysis as existing modified Reynolds equations using the slip-flow models. Numerical calculations of load carrying capacity are performed for comparison with those obtained from other modified Reynolds equations. It is found that the present modified Reynolds equation provides intermediate characteristics between those from Burgdorfer's modified Reynolds equation and those from the generalized Reynolds type lubrication equation based on the Boltzmann equation.

抄録全体を表示

The overall objective of this study is to introduce a wear map of ceramics which shows the regions of dominant wear modes obtained in a wide range of materials and operating conditions by modeling. In this report, wear modes of various ceramic materials observed under a wide range of experimental conditions are classified experimentally. For this purpose, friction and wear tests are carried out on three kinds of ceramics sliding against themselves under various contact pressures, sliding velocities and temperatures. Sliding wear phenomena of ceramics observed in all tests can be classified into two types called "mild wear" and "severe wear" based on the ratio of worn surface roughness, R_y, to mean grain size, D_g. Specific wear amount is always less than 10^-6 mm³/Nm when worn surface roughness becomes relatively smooth (R_y/D_g<0.2) compared with mean grain size in the mild wear region as a result of running in, nevertheless, relatively rough wear surface (R_y/D_g>0.5) in the severe wear region gives a wear rate larger than 10^-6 mm³/Nm. The transition from the mild wear to the severe wear occurs when normal load, sliding velocity or temperature attain a certain value. The quantitative region of each wear mode is clearly shown in the load-sliding velocity diagram and load-temperature diagram, experimentally.

抄録全体を表示

Ball parts assembled in rolling bearings are finished by the lapping process. Usually, the ball parts are lapped while rolling between coaxial circular grooves machined on the two laps. These grooves have a V shape or a semicircular cross section. The difference between radius of the semicircular cross-sectional grooves and the radius of the ball parts is very small. The high conformity, which stands for degree of agreement between the radius of the balls and the grooves, lead to a high sphericity of the ball parts. However, in the case of lapping with a large stock allowance, a form error which has three lobes along the circumference of the cross section increases. In this report, the lapping condition and mechanism of increase in the form error with three lobes are experimentally investigated. A small decrease in conformity between the ball parts and the grooves results in an increase in the form error. In this condition, gap between the ball and groove becomes enough large for a form error of the ball to give rise to displacement of the ball mass center.

抄録全体を表示

In this paper we propose a simple evaluation method for form deviation of complex planar curve contours based on coordinate measurement data. Using this method, the contour curve is expressed in terms of its optimum interpolation circular arcs, whose interpolation accuracy can be specified arbitrarily, and the form deviation is defined as the deviation between the points of measurement and the arcs. The geometrical form deviation of the profile is evaluated using the least squares method. It has been demonstrated that the difference between the deviation definitions, relative to the contour curve or to the arcs, is not larger than the circular interpolation accuracy. In other words, the accuracy of the evaluation can be sufficiently guaranteed. The method is easy and convenient for engineering applications. An actual evaluation example for a planar cam profile is shown. The result confirms the validity of the proposed method.

抄録全体を表示

One-way clutches are examples of machine elements that apply the mechanism of self locking and are widely used in various power transmission systems. A study on the characteristics of self-locking can not proceed without regard for the effect of the deformation caused by the stress occuring at the contact surfaces between sprags and the races of inner and outer wheel. In this paper, the characteristics of self-locking is studied by using the finite element method. It is shown that a statically simple relation does not exist in the actual self-locking condition and slipping takes place over some part of the contact region.

抄録全体を表示

A precise analysis has been made for a design method in order to obtain the profile of a basic thread surface of a hob which cuts a given screw rotor surface, especially in the case in which a screw thread profile is given as a series of dispersive points. Firstly, it is clarified that under certain circumstances a singular point (a cusp) will occur on the basic thread profile of the hob corresponding to the screw rotor profile, and that, because of the cusp, the interference phenomenon will occur on the basic thread surface of the hob. Secondly, numerical analyses are proposed to determine the interference by using a conditional equation of the cusp as well as to obtain the basic thread surface of the hob. Then, these analyses are theoretically verified by applying them to the dimensions of a pair of screw rotors.

抄録全体を表示

A precise analysis has been made for a design method in order to obtain the profile of a relief grinding wheel surface for a hob which cuts a given screw rotor surface, especially in the case in which the basic thread profile of the hob is given as a series of dispersive points. Firstly, it is clarified that under certain circumstances a singular point (a cusp) will occur on the surface of the relief grinding wheel corresponding to the basic thread profile of the hob, and that, because of the cusp, the interference phenomenon will occur on the surface of the relief grinding wheel. Secondly, numerical analyses are proposed to determine the interference by using a conditional equation of the cusp as well as to obtain the profile of the relief grinding wheel surface. Then, these analyses are theoretically verified by applying them to the dimensions of a pair of screw rotor hobs.

抄録全体を表示