Search

学位の種別: 課程博士, 審査委員会委員 : （主査）東京大学教授 一條 秀憲, 東京大学教授 後藤 由季子, 東京大学教授 堀 昌平, 東京大学准教授 瀬戸口 留可, 東京大学特任講師 畠 星治

article

This paper presents a robot system for handling various objects in home or office environments. A fixed manipulator utilizes marks on objects. A mark consists of two parts. One is the outer part, which indicates its 3 D pose, and the other is a QR code, which is stored information of an object (e.g. its size). A QR code is a kind of 2 D barcode. A user attaches several marks on objects. The manipulator accesses the information using a camera on its hand. The manipulator estimates pose of objects measuring marks and using least-squares method. Then, the robot can handle and convey the objects through planning a sequence of conveyance and a pose for grasp that is based on the pose of objects, candidates for grasp pose informed by the mark, and signals from proximity sensors. Experiments are conducted to verify the whole system.