Computable Minds - www.computableminds.com
57

# Robots, Kinect and the SLAM problem

Posted on: Dec, 14th 2010
In one of the posts where I explained the how Kinect works, I said that wouldn't be strange that it will be use to build robots. Now are emerging the first robotic machines that someday will rebel and will begin to kill humans. Here I talk about some of them.

As a stereographic camera for investigation projects can cost about \$ 2.300 and Kinect, being a lot of more cheaper than a stereographic camera, does the same, is logic that begins to emerge all kind of investigation projects to take advantage of it to do robots. Also, would be a good device to the help the blind people to walk, translating their enviroment to acustical signals with Kinect. Next I explain how are some of the robots that are arising.

The first is mounted over a kind of aspirator robot. It's realized by a Phd MIT student, called Philipp Robbel. The robot can recognize the environment, where is located itself in it, identify humans and obey to their gestures.

Is interesting explain this robot in depth. To create a 3D map of its surroundings sends, through of a Wifi connection, the information obtained by Kinect to a computer that works as the brain of the robot. This executes an algorithm called SLAM (Simultaneous Localization And Mapping) that creates a map of the enviroment at the same time that does a tracking of the actual position of the robot, with the intention of that the robot can know where it is, how is the world and in consequence where to go. Is a complex problem, by one hand we can't trace the position of the robot if we don't know the map, and on the other hand we can't build the map if we don't know what is the position where the robot is capturing images. The accumulated errors at every step do to the robot more difficult estimate its position in the space and calculate the map without distortion, being this one of the problems that remain still to resolve in artificial intelligence.

The next robot is a quadricopter capable of fly chasing a marks without crash with nothing.

Following, a four wheels robot that can displace very fast without crash:

Finally, a robot that don't use Kinect directly. In this case is the human the one that use the Kinect as interface to control the robot, a very interesting option to control robots in the distance in dangerous situations, although still the project isn't very advanced.

Categories: , , ,

Share:

Copy and paste in your page:

Share

To check if you are human answer the question correctly:

None of these data will be stored.

(Write the e-mail)

Required field.

(Write the e-mail or several e-mails separated by coma)

Required field.

To check if you are human answer the question correctly: