Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : May 10, 2017 - May 13, 2017
We have developed a tracked mobile robot with multiple robotic legs which can work by itself in a disaster area. It can perform handling tasks such as object gripping and carrying by the use of two fore legs as robotic arms. To reduce burden of remote control on an operator for such tasks, this study presents a method for detecting gripping positions of an object for the two arms using an RGB-D sensor. The method is based on plane detection of a polyhedron object. Plane information of an object is detected by the depth sensor, and two surfaces having opposing normal are selected as a candidate for planes to be gripped. Joint angles of both arms for gripping the planes are calculated with position and posture of the robot using inverse kinematics. If the robot can grip the planes by both arms, they are selected as the gripping surfaces, then the robot moves to the obtained gripping position and performs gripping and lifting up the object. Several experimental results verified the effectiveness of this method.
