[in Japanese]

View full abstract

[in Japanese]

View full abstract

In construction and maintenance of steel bridges, it is necessary to develop the method that can infer axial force of high-strength bolts in joints by nondestructive inspection. The loosening of high-strength bolts has to be inspected by experts with hammers in the past. The axial forces of bolts have to be reasoned quantitative for the accurate inspection. We have develolped a system for inferring axial force of high-strength in mediuam and small-sized steel bridges using waveform data provided by a hitting of automatic looseness detector. The system has been based on a neural network with faculty of pattern recognition. However, the system was not able to be applied long high-strength bolts on the joints of great length bridges. In the present paper, the system was constructed for inferring axial force of high-strength bolts jointed in not only medium and small-sized bridges but also great length bridges and jointed in secondary members. The system can infer axial force simply and accurately in the building site. Firstly, the new method for adjusting the automatic hammer was decided and the usefulness was verified. Secondly, the relation between grip length of high-strength bolts and the inferred axial force was cleared in details, because the grip length is the most concerned bridge scale. The practical system was constructed in consideration of grip length. Axial force of various high-strength bolts on the joints was inferred using experimental model. The experimental model was made after I-girders of a steel bridge, given the same structure and almost same dimensions. The system was verified for practical use in the case of various bolts on the secondary members. The system can infer both the installed axial force in construction and the residual axial force in service. Using this system, axial force of bolts can infer easily without extracting high-strength bolts. Therefore, improvement of productivity is anticipated in construction and maintenance of steel bridges.

View full abstract

Conventional computing which processes information, such as artificial intelligence, and natural language processing which is computing of language, have been perceived individual processing. However, human intelligence is supported by the functions of language, if we wish to realize human intelligence on a computer, we have to fuse these computing and achieve a framework for computing in which any information is processed with language. Based on this idea, the concept of everyday language computing has been proposed. In this paper, for a concrete example of everday language computing, we develop a dialogue system for travel consultation which interprets a user's intentions by understanding the meaning produced by words in a consultation dialogue, and then we show the process of how travel information required by a user will be retrieved based on the meanings the system obtained through the dialogue. As a framework for processing the meaning of a dialogue, we adopt the framework for a meaning base, which is defined in systemic functional linguistic theory as a human experience model constructed based on the linguistic system. And we analyze and generate meaning of speech by applying the meaning base to travel consultation, and try information processing with the meaning of speech. And then we show the practicability of our proposing method through the dialogue simulation system.

View full abstract

An array antenna, which has a large number of radiating element, is divided into subarrays. The current distribution of the planar array antenna is approximated as a Taylor distribution using the genetic algorithm to combine subarrays having the same current distribution. Doing this simplifies the feed systems. By considering the planar array antenna as an aggregate of subarrays, designs, production, inspection, and maintenance are simplified, which in turn reduces costs.

View full abstract

the objective of the elevator group control system is to efficiently operate an elevator system composed of limited number of elevetor cars, and to assign an appropriate elevator car to an issued hall-call. There are several elevator group control systems such as minimum long-wait algorithm and area-based control algorithm. These algorithms are desingned to reduce service time(waiting time + traveling time)by decreasing waiting time. This article proposes a new algorithm called total fuzzy control algorithm, which focuses not only on waiting time but also on traveling time. In other words, this algorithm is designed for reducing traveling time in addition to waiting time reduction. The system proposed in this paper uses fuzzy expert system. That means the inference potion of the expert system uses fuzzy inferences. The study proceeds step-by-step beginning from fuzzy expert system without traveling time consideration and behavior of fuzzy control system is investigated, and then using this base the total fuzzy control system that includes the consideration of traveling time is constructed. Effectiveness of the total fuzzy control algorithm is thoroughly studied by computer simulation and the results show to give better service in some specific traffic conditions.

View full abstract

Linear discriminant analysis by R.A.Fisher is based on the single linear projection of the whole data set, that provides a convenient index for classification. If the data set to be classified has a region where the data of the different classes overlap and has nonlinear decision boundaries, Fisher's linear discriminant analysis may not be able to demonstrate good classification performance. In this paper, we propose a new linear discriminant approach(switching linear discriminant)which derives the linear discriminant function for each cluster obtained by a fuzzy clustering algorithm which takes local nonlinearity into consideration. The fuzzy clustering algorithm locates cluster centers around the nonlinear boundaries for the switching linear discriminant, which thus provides several indices for classification in a form of simplified rules. The effectiveness of our method is shown in the numerical examples.

View full abstract

This paper deals with Multiple Traveling Salesman problems(MTSP), which is one kind of combinatorial optimization problems. Unlike the conventional TSP, in MTSP plural salesmen can make travels to minimize a given cost function. In these problems, to shorten total tour length and to even out the tourlength among the salesmen are simultaneously required. A solution of an MTSP forms N clusters upon the given cities. the basic idea of this paper is that the initial N clusters are not determined strictly, but defined roughly. Final travering plan is formed step by step in consideration of the balance of length of all local tours. This paper uses the fuzzy c-means method to make the rough clusters. the fuzzy c-means method is one of representive methods of fuzzy clustering. Proposed procedure can easily cope with the k-TSP, which has common starting and returning point of MTSP. The proposed method can be considered and important application of fuzzy clustering method.

View full abstract