View Full Version : Summary of replies on: robot arms, turn tables

06-18-2007, 09:02 PM
Once again, thank you all for your contribution.


Dear All,

We are looking for information on different types of robot arms and/or turn tables that can be used to manipulate in space with rigidly attached pair of cameras+projector in order to image body from different views and consequently find full 3D shape of it. What we expect to see from such a robot arm on its output interface is information about change in attitude/pose between two successive positions (camera views) in space, i.e. three translation and three rotational parameters. As it is probably obvious, we need those to easy up problem of range image registration. Alternatively, one could perhaps always try and build up such robot arm (or mentioned turn table in simplest case for smaller object of size and complexity where the object can be put on the turn table and cameras kept stationary) on his/her own. However, we would like to explore first perhaps some readily available solutions on Web. For example http://www.reisrobotics.com/, although that may turn out to be as overkill, in terms of size, installation, accompanying utility for manipulation; for a 'simple' goal what we need from that robot arm: translation + rotation.

Thank your for your output, summary will follow.

Best regards,

Dr.sc. Tomislav Pribanic
Faculty of Electrical Engineering and Computing
3 Unska
10000 Zagreb
Tel.: +385 1 6129 867
Fax.: +385 1 6129 652
E-mail: tomislav.pribanic@fer.hr


We have just completed the realization of a robotic arm. It will be displayed on the lab WEBpage soon. We may discuss in the future

Alberto Borghese borghese@dsi.unimi.it


I don't know the particulars of your application, but you may look at some off-the-shelf solutions that already exist like the T-Ring from Ohio Willow Wood:

Or a standard scanner:



Or outsource it to a specialist:


If that doesn't fit your application (or if you want to build it yourself), please let me know. (I am a roboticist by training.) What you are describing is a general 6DOF robot which can be performed in a variety of ways (serial, parallel) and by a variety of manufacturers (ABB, Fanuc, Staubli, Motoman, Panasonic, etc.). The main things to worry about when specifying a robot is payload weight and size and workspace requirements.



Philip A. Voglewede
Assistant Professor
Department of Mechanical Engineering
University of South Carolina
300 Main Street, Room A219
Columbia, SC 29208
email: voglewed@engr.sc.edu


Dear Tomislav,

To get the best out of available space and attain the greatest accuracy from the robot, a consideration is to match your workspace geometry with that of the robot. This website shows the main configurations:
http://www.ifr.org/pictureGallery/robType.htm. If the body is lying down the immediate answer is something similar to a CT scanner. Or if upright, as you suggest - a turntable and horizontal scanner. Both could be produced with a Cartesian frame as your starting point, the components are relatively
standard: http://www.techno-isel.com/tic/LinMoProducts.htm#XYTABLES.

An emphasis on automated factory layout has always been to ensure people stay out of the workspace, which is obviously not the case for surgical robots. I've often wondered how this safety issue was tackled with systems such as the Da Vinci robot. Has anyone experience of this?


Graham Webb
Trainee Clinical Scientist
Department of Medical Physics and Bioengineering, St Georges Hospital LONDON