updated 8/5/2011 11:19:12 AM ET 2011-08-05T15:19:12

Scientists at the Tokyo Institute of Technology have demonstrated a robot that has tackled the next big problem in robotics: what to do when a solution isn't part of its programming.

"So far, robots, including industrial robots, have been able to do specific tasks quickly and accurately. But if their environment changes slightly, robots like that can't respond," said Osamu Hasegawa, associate professor with the Hasegawa Group at the Tokyo Institute of Technology (see the video above).

A truly autonomous robot will need to adapt to new situations and learn from past experience, which is exactly what this robot is intended to do.

"This robot remembers only basic knowledge, and it can apply that knowledge to its immediate situation. If it doesn't know enough, it stops, and reacts by saying, 'I can't do this because I don't know how.' So, if you teach this robot just the things that it can't do, it incorporates those things as new knowledge, and it can solve the problem overall, by including that knowledge," Hasegawa said.

Using a technology called SOINN (Self-Organizing Incremental Neural Network) the robot can make an educated guess for solving a new problem based on what it has learned previously. In the demonstration, the robot had to figure out how to pour water in a cup (tiny plastic balls were substituted to prevent potential water spills around sensitive electronics) and make the water cold using a block representing ice.

Even small things, such as the need to set down the bottle before the ice cube can be picked up, are all intuitively decided by the robot based on what it has learned, rather than specifically programmed tasks.

The robot doesn't just learn from experience. If it cannot figure out a solution to a task, it can ask for help and then remember it the next time it encounters the problem. Hasegawa also envisions robots connected through the Internet that can share solutions with other robots when they discover how to overcome a problem, meaning the robots can all learn together more quickly than they can alone.

© 2012 TechNewsDaily


Discussion comments


Most active discussions

  1. votes comments
  2. votes comments
  3. votes comments
  4. votes comments