Image from http://www.dailymail.co.uk/sciencetech/article-2626598/Catch-Watch-incredible-robot-arm-never-drops-thing-set-blasted-orbit-grab-space-junk.html
Previous methods to accomplish such tasks was to programme the arm to calculate the trajectories of the objects. However, this method takes too much time when a decision has to be made in a split second. Thus, inspired by how we learn to catch flying objects ourselves, the researchers taught the robotic arm through imitation and trial and error. This means that they taught it how to move by first throwing various objects, such as empty and half full water bottles and tennis rackets, at it and manually controlling the arm to catch them. These particular objects were chosen because their centers of gravity were not necessarily coincidental with their actual center, allowring the robot to learn to handle various situations.
The arm already has many practical applications, one of it being in space, where it can be used to retrieve space debris. However, I would personally like to see it be used as a prosthetic arm. This would definitely be a great gain for those who have lost their arms. One problem though, is that would the arm be under the users control? Or would it just catch or dodge, even if the user did not see it coming? The future possibilities are just so exciting for this new invention!
Another point I would like to make is that would this become the starting point for the robot takeover of Earth? (Haha!) Although I may say this jokingly now, but the main aspect that separates is from robots is our ability to learn new things independently and apply them in novel situations, whereas conventional robots had to be pre-programmed and told what to do. Thus, today it may start with an arm learning how to catch but tomorrow, it may be a robot learning how to create weapons (which can also be done through imitation). Do you think it is possible? Comment down below!
You may also like to watch the video here: http://m.phys.org/news/2014-05-ultra-fast-bionic-arm-video.html
Reference:
Seungsu Kim, Ashwini Shukla, Aude Billard. Catching Objects in Flight. IEEE Transactions on Robotics, 2014; 1 DOI: 10.1109/TRO.2014.2316022
No comments:
Post a Comment