IROS 2015 Paper Abstract

Close

Paper ThAP.6

Miyazaki, Tetsuro (Yokohama National University), Kanekiyo, Akihiro (Yokohama National University), Tsuchiyama, Yutaka (Yokohama National University), Sanada, Kazushi (Yokohama National University)

Experimental Validation of Integrated Robot Design Using Ball Throwing Robot

Scheduled for presentation during the Poster session "Late Breaking Posters" (ThAP), Thursday, October 1, 2015, 09:45−10:00, Saal G1

2015 IEEE/RSJ International Conference on Intelligent Robots and Systems, Sept 28 - Oct 03, 2015, Congress Center Hamburg, Hamburg, Germany

This information is tentative and subject to change. Compiled on July 20, 2019

Keywords Integrated Task and Motion Planning, Motion Planning for Manipulators, Redundant Robots

Abstract

This poster proposes an integrated robot design method which designs a robot body and a robot motion simultaneously to maximize robot motion performance of a target task. Different tasks require different appropriate bodies, for example, humanoid robots have human-like bodies to execute versatile tasks of a human. On the other hand, industrial robots have specialized bodies to execute limited tasks efficiently. In case that the target task is limited, the integrated robot design method is effective to improve the robot motion performance. The integrated robot design method is available to design a multi degrees of freedom (DOFs) robot, and it is validated by ball throwing experiments using 9 DOFs arm robot. Design parameters are a motion pattern and robot body parameters, and these initial values are optimized to maximize (or minimize) evaluation functions which are given from ball throwing task conditions. The ball throwing task conditions contain (i) maximization of ball flying distance, (ii) collision avoidance between flying ball and robot hand, (iii) velocity evaluation of motion, (iv) self-collision avoidance and (v) limitations of joint angle, angular velocity and torque. The robot motion performance of the ball throwing task is defined by the ball flying distance, and it will be maximized by using condition (i). Other conditions are utilized as the constraints to obtain the realizable robot motion. A structure of 9 DOFs arm robot link system is defined as that is able to realize a ball throwing motion like a human. Therefore, a ball throwing motion performed by a human is measured by an optical motion capture system, and its joint angle trajectory is utilized as an initial value of robot motion pattern. The initial motion pattern is transformed to a motion pattern of the robot, and the robot body parameters are designed simultaneously. Link lengths are defined as the body design parameters, and other body parameters will change along with the link lengths. Link mass, center of gravity and moment of inertia are represented as functions of link lengths. Steepest descent method is utilized as the calculation algorithm of the design parameters. The evaluation functions related with the ball throwing task conditions are differentiated with respect to the design parameters in the recursive algorithm, and these gradients are used to update the design parameters. Calculation results are verified by experiments using the real robot system. 2 cases, which contain improved motion 1 (the robot motion pattern is only designed) and improved motion 2 (the robot motion pattern and the robot body parameters are designed simultaneously), are obtained, and these average ball flying distances of 5 trials are 1.33 m and 1.87 m. As a result, the ball flying distance of improved motion 2 is longer than that of improved motion 1, and effectiveness of the integrated robot design method is demonstrated.

 

 

Technical Content © IEEE Robotics & Automation Society


This site is protected by copyright and trademark laws under US and International law.
All rights reserved. © 2002-2019 PaperCept, Inc.
Page generated 2019-07-20  00:26:24 PST  Terms of use