Monday 23 April 2007

A Chess Playing Robot

For my day job I work at the Australian National University in the area of Robotics. This mainly involves writing software that allows robots to operate, navigate, and appear more intelligent in interacting with humans. I've included a couple of pictures of the robot I am mainly working on for Questacon.
One of the possible outcomes of my work (but not necessarily with this project) is a chess playing robot. By that I mean a robot that is developed enough to enter and compete in a chess tournament autonomously. While for non-roboticists this may seem some way off, most of the building blocks are already in place.
The robot in the picture was designed by David Austin and he also wrote most of the initial software. Included in the software for the robot is a chess program, although it is hardly complete at the moment. The robot also has a vision system, allowing it to keep track of the game via the stereo cameras on its head. It also has a speech recognition and response system which allows humans to converse with it, and receive sensible answers. This is particularly important as one ANU Professor (unnamed) who rejects the notion of computer chess as AI stated that "You may be able to develop a program to beat Gary Kasparov, but it isn't able to discuss the game with him after it's over"

And finally it is a mobile robot, which means it can navigate sensibly, as well as being able to avoid running over things. All that is missing are robot arms, which would allow it to move it's own pieces. However there have been separate projects that do exactly that, so implementing that feature is also possible.
What this means that all that is needed to build a real chess playing robot is the time and energy to glue it all together. And personally I wouldn't be surprised if that happens within the next 3-5 years.

No comments: