Understanding of Geometric and Mechanical Information Based on Multi-fingered Hand Manipulation with High-speed Vision

2014 Fiscal Year Final Research Report

• Project/Area Number: 25540113
• Research Category: Grant-in-Aid for Challenging Exploratory Research
• Allocation Type: Multi-year Fund
• Research Field: Intelligent robotics
• Research Institution: The University of Tokyo
• Principal Investigator: SENOO Taku 東京大学, 情報理工学(系)研究科, 助教 (10512113)
• Project Period (FY): 2013-04-01 – 2015-03-31
• Keywords: ロボット工学 / 画像処理 / 多指ハンド / 力学同定

Outline of Final Research Achievements

A method of simultaneously identifying geometric and mechanical information of an object was proposed using the data set of time-series multi-view image captured with high-speed vision. By actively manipulating a measurement objet to get multi-view images with a single camera, the 3-D shape reconstruction was performed in voxel space on the basis of a space carving method using the objet silhouette, and the inertia moment and mass distribution were also estimated in fitting the object movement to the mechanical constraint such as the conservation of angular momentum and the conservation of energy.

Free Research Field

知能ロボティクス