Kohei Nomoto

Similarity measures using fuzzy thesauri for case-based reasoning

The conventional similarity measure between a query and a case is measured by counting the number of indices in which values of the case are the same to those of the query. The values of an index, however, are not exclusive of each other, and hence, even if a value of the case is not the same to that of the query, they may sometimes have a relation. The paper presents new similarity measures which consider the degree of the relations among the values using fuzzy thesauri.

Target tracking algorithm in radar reference coordinates using coupled filters

This paper proposes a tracking algorithm in radar reference coordinates (RRC) using coupled filters. Tracking filters are discussed from the viewpoint of the fundamental problems of both modeling accuracy and real-time computational requirements. In most cases, if one becomes better, another gets worse. A trade-off between them always exists. Then our algorithm satisfies these conditions as much as possible by use of coupled filters. Through several simulations, validity of our algorithm has been confirmed in tracking non-maneuvering target.

Attention allocation model for manufacturing skill

The purpose of this paper is to present an analytical model of skill of a worker who operates an automated manufacturing machine. The skill is the way of paying his attention while the machine runs for the first time by a new program. In such a case, he should watch the process to find out an unexpected motion caused by a program mistake, and if it is detected, he will interrupt the program-mode immediately before it results in a serious accident.