感覚器官をもった人工の指による物体の認識

DOI Web Site オープンアクセス

書誌事項

タイトル別名
  • Pattern Recognition of an Object by Artificial Fingers with Sense Organs

この論文をさがす

説明

In the recognition of the nature of an object by artificial fingers, the selection of sense elements and the processing method of the information from them are important. In this study a sense element using a strain gauge advanced recognition than that using on-off was made and tested to get the sense information in the form of analogue voltage for more switches. This sense element is superior in line arity and bears repetition of the load. It is seen that this element can get information necessary to classify the shape, size and material of the objects.<br>The concept of “attributes” to describe the nature of the object is suggested, and a method of classification of each attribute is also presented. It is easy to add the new attribute, and the amount of the calculation for classification is reduced in this method. The linear discriminant function method is one of the most popular recognition theories. On the other hand there is a theory using the Bayes minimum risk decision function, and this can be translated to a Chebychev polynomial discriminant function. We compare both methods using the same data. We could not find significant differences between the two methods in the correct rate of answers, but the latter is preferable from the point of view of the time for calculation.

収録刊行物

詳細情報 詳細情報について

問題の指摘

ページトップへ