2018 年 84 巻 1 号 p. 70-76
To grip various objects indoors by a robot, it was investigated how many kinds of object on a desk are handled and how they are gripped in a human life, in which a subject with wearable camera records images for 30 days. It was proven that not more than hundred kinds of objects are handled in gripping ways corresponding to the shape and size of each object. They were classified to 9 categories. Hands of human subject were restricted so that they only act like parallel grippers. According to the gripping way of restricted hands, categorization was revised to 12 categories. To cope with all categorized objects, the gripping methods were proposed for dual robot arms equipped with parallel grippers. It was preliminarily proven that the proposed method has potential for practical robot arms to grip all categories of objects mentioned above.
