[in Japanese]

View full abstract

In this paper, we propose the face identification method which is robust for lighting based on the feature points method. First of all, the proposed method extracts an edge of facial feature. Then, by the hough transform, it determines ellipse parameters of each facial feature from the extracted edge. Finally, proposed method performs the face identification by using parameters. Even if face image is taken under various lighting condition, it is easy to extract the facial feature edge. Moreover, it is possible to extract a subject even if the object has not appeared enough because this method extracts approximately the parameters by the hough transformation. Therefore, proposed method is robust for the lighting condition compared with conventional method. In order to show the effectiveness of the proposed method, computer simulations are done by using the real images.

View full abstract

In this research, we propose the system which discriminates a vowel only by operation comparatively simple for mobile apparatus. Mobile apparatus, such as a cellular phone, is small. It is difficult in hardware of small mobile apparatus to apply the speech recognition system currently developed. Then, we extract the features from a vowel amplitude waveform directly, and vowel recognition is performed by the Mahalanobis distance. Finally, in order to illustrate the effectiveness of the proposed method, computer simulations are done.

View full abstract

In this paper, an EEG control system is constructed. In this system, the properties of music are adjusted automatically with EEG, and EEG is controlled by listening to that music. EEG is used for human healing music in sleeping, relaxing, and so on. EEG has complex wave form. Therefore, we think it is important to analyze EEG in this system. In this paper, we recognize the EEG pattern by using factor analysis and neural networks. Factor analysis is used for extracting the characteristics of EEG. NN is used for learning the characteristics of EEG. Then teacher signal data of NN uses the data of the characteristic of the music. Furthermore, in order to demonstrate the effectiveness of the proposed method, we show simulation examples.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

In this paper, we look back upon the circumstances from the Hanshin Awaji great earthquake to the project called Dai-Dai-Toku sponsored by Ministry of Education, Culture, Sports, Science and Technology, and the following points are considered. (i)Why is it that there was no motion before the earthquake disaster? (ii)Why we took 7 years to realize the project. (iii)What were the main reasons for the realization?

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

This paper discusses a configuration of a fundamental robot and some applications of robot. We will forcus the problems of robot in future at the standpoint of control engineering. In future, robot technology makes a new system with an information technology and an information network. We point out some problems of this system.

View full abstract

Today’s most important issues related to automotive traffic are traffic accidents and environmental concerns. Automotive technology has continued to be improved in order to solve these problems. New automotive mechanisms developed in this context are very usually planed on the assumption that they are controlled using microprocessors. On the other hand, Intelligent Transportation Systems, which contain many automotive control systems, are now under the construction aiming for high efficiency and safety of the automotive traffic. This paper describes current research trend of vehicle control and consider how tomorrow’s automobiles will be from a viewpoint of control technology.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

In this paper, we present a method by which different facial expressions can be created from an expressionless face image. In a visual scene our method detects and extracts the position of a face using a neural network based method. The required facial expression can then be created by selecting it from a list of keywords, for example smiling, angry, surprised etc. The user of our method is required to provide only the image and the desired facial expression. The rest of the processes are fully automatic including the selection and detection of the control points necessary to perform the warping. Our warping algorithm is based on the triangulation method that uses triangles to perform warping.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

The use of soft computing techniques such as neural networks(NNs), genetic algorithms(GAs), etc. in the financial market has recently become one of the most exciting and promising application areas. In this paper, we propose two kinds of decision support systems(DDSs) for dealing stocks which utilize NNs & GAs. In the first half of this paper, we propose the DSS which improves the traditional technical analysis. GAs are utilized in order to find effective ways of dealing stocks using the proposed DSS. In the latter half, we propose the DSS which exploits the forecasts of stock price up to four weeks in the future given by NNs. Several computer simulation results confirm the effectiveness of the proposed DSSs.

View full abstract

[in Japanese]

View full abstract

Vision is a sensing organ which acquires 80-90% of various information in space. Therefore, the visually impaired person makes index of action the information which should be acquired from vision depending on other sense organs and auxiliary apparatus. Although various researches for independent walk realization of a visually impaired person have so far been made, there are few systems which left for the user side. Then, in this research, in order to take in a user's view-point and a visually impaired person's demand, the communication of information system which asked cooperation of the visually-impaired school and left for the user side and which is used together with a while cane is proposed.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

Hybrid position/force controller for the tendon driven robot was proposed. Combining the joint space orthogonalization method by Liu et al.[1] and the tendon driven robotic mechanisms controller by Kobayashi et al.[2], we constructed the controller and showed its stability, Simulation and experiment showed the efficient of controller.

View full abstract

Tendon-driven robot arms are very suitable for position/force Hybrid control, because we can adjust the joint stiffness semiactively. C.I.Park,et al,[3] proposed a position/force controller by combining the joint-space orthogonalization method given by Y.H.Liu,et al.[1] and the controller for tendon-driven robotic mechanism given by H.Kobayashi,et al.[2]. In this paper, their results is extended to an ANN controller, and a set of sufficient conditions for the UUB is obtained.

View full abstract

This paper proposes a parallel computation algorithm for very highspeed active contour. The algorithm is implemented on the visionchip which has SIMD architecture. The performance is evaluated byreal time tracking of high speed finger motion.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

We have been developing a "muscle suit" that provides muscular support to the paralyzed or those otherwise unable to move unaided, as well as to manual workers. The muscle suit is a garment without a metal frame and uses a McKibben actuator driven by compressed air. Because actuators are sewn into the garment, no metal frame is needed, making the muscle suit very light and cheap. Most medical and welfare apparatuses have been developed for the caregiver rather than for the patient in need of nursing, detracting from the patient’s sense of self-sufficiency and generally discouraging the patient. With the muscle suit, the patient can willfully control his or her movement. The muscle suit is very helpful for both muscular and emotional support. In this paper, we propose an armor-type muscle suit in order to overcome issues of a prototype system and then show how abduction motion which we believe is the most difficult motion for the upper body, is realized.

View full abstract

We have been requiring the human assist devices for elderly people to support works in daily life. In order to develop these devices which come in contact with human bodies, these devices must have hybrid element which is constructed with passive elements to change shape flexibly by a human body and active elements to control shape and output force. Thus we propose “Silicon Outer Fence Mold Actuator” which is driven by low-pressure. In this paper, we describe the force control performance and repetition accuracy of the proposed actuator.

View full abstract

[in Japanese]

View full abstract

We have been studying exoskeleton type robots to assist the motion of physically weak persons such as elderly persons or slightly disabled persons in daily life. In this paper, we proposed a 3DOF exoskeleton type robot to assist the upper-limb motion (shoulder joint motion and elbow joint motion) of physically weak persons. The proposed robot automatically assists the human motion mainly based on the skin surface electromyogram (EMG) signals. Fuzzy-neuro control has been applied to realize the sophisticated real-time control of the exosk-eleton type robot. Experiment has been performed to evaluate the proposed exoskeleton type robot.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

The paper proposes an effective non-destructive inspection method using an electro-magnetic wave radar which detects the defects such as cracks or cavities in concrete structures and measures their positions and thicknesses (for cavities). This is achieved by a pattern matching between actual and predicted reflected waves based on a signal propagation model.

View full abstract