Online from: 1973
Subject Area: Mechanical & Materials Engineering
Options: To add Favourites and Table of Contents Alerts please take a Emerald profile
Downloads: The fulltext of this document has been downloaded 405 times since 2009
Article citation: Clive Loughlin, (2009) "Fun and games", Industrial Robot: An International Journal, Vol. 36 Iss: 2, pp. -
Our Viewpoint in this issue, “Robotic issues on vision-based human motion analysis” by Dr Honghai Liu from University of Portsmouth raises some interesting possibilities.
If you write a program for one type of robot, for example a SCARA, it is then not too challenging to modify the program so that the same function can be achieved either by a different-sized robot of the same basic type or a robot with completely different kinematics, for example an anthropomorphic robot.
One of the major challenges still remaining for robotics is still that of rapid and robust programming. There are lots of potential applications for robots for tasks that are still being performed by people, and a major reason is often simply that the effort required to install and program a robot to perform a task that may only be required for a day or two before a major or minor adaptation is required.
If we could develop a way to generate a robot program by analysing human motions as the task is performed manually, then the same sort of transformation matrices that we used to convert a program for one robot so it can be performed by another, can be used to transfer the program from a person to a robot (anthropomorphic or otherwise).
Of course, this is a very challenging task that is several orders of magnitude more difficult than a robot-to-robot transformation. Whereas, Robot A has joints of largely a fixed length (temperature and loadings will of course, have an impact), and axis of rotation with a defined center point, which can be fed into a transformation matrix to create the program for Robot B; the same cannot be said of a person.
The joints in our arms each typically have 3 degrees of freedom whereas those of most robots just have one. And the surfaces of the ball and socket joints are perhaps roughly what you might call spherical but with significant natural variations.
However, the potential benefits of solving these challenging problems make the journey worthwhile. As with many useful industrial developments, the basic research often gets done by companies with little or no interest in industrial automation. One example of this is very topical as I write this Editorial about 6 week before Christmas.
The shops are full to bursting point with Wii games and the remote controls that allow your movements to be analysed and used to animate your current alter ego as it fights Star Wars type battles on the screen in front of you.
The market for the software was in toys, and because of this, the 3D motion analysers that would have set you back several thousand dollars are suddenly being literally given away.
If the technology can be developed sufficiently to allow you to pick up your light saber then it is not a great leap of faith to see that it could find application on the shop floor.