Keepon is a small yellow robot that displays basic gestures in order to interact with humans, especially children. It was developed by Prof. Hideki Kozima of the National Institute of Information and Communications Technology(NICT), Japan.

The main aim of the Keepon project is to find out how to improve people's interaction with a robot. People usually feel uncomfortable around robots and children tend to get scared; hence a friendly looking robot like Keepon was created. To further improve the interaction, Keepon is programmed to videotape the reactions of children which are later analyzed. This helps scientists think of new ways to help people feel more at ease while interacting with robots. Keepon is also used to treat children with autism by making Keepon interact with them.

Because of its ability to nod(+/-40 degrees), shake(+/-180 degrees), rock sideways(+/-25 degrees) and bob(+/-15mm stroke), Keepon is able to  interact with people. Keepon has four motors, its eyes are the cameras, and its nose is a microphone. Because of its ability to make eye contact and joint attention, Keepon is used to study social interaction among humans.

To identify a human, Keepon takes real-time image streams (30 frames/sec) using the camera and identifies humans based on their skin color and average face templates. It maintains joint attention by first generating several of its own hypotheses and then based on the images from the camera, it tries to identify the direction of human attention and on identifying the direction, the four motors of Keepon enable it to face the human.

Michalowski's faculty advisor Professor Reid Simmons deals with research on Human-Human non verbal interaction says, "Rhythm is very important in human-human interaction, so it is not surprising that it is so compelling in human-robot interaction."  Non-Verbal interaction basically refers to gestures of nodding heads and other movements which are reactions to other people's gestures and show that we are able to understand the other person. Similarly, Keepon uses rhythms to react to humans and are able to react in sync with human speech which makes those interactions look more natural. By reacting to various rhythms and beats, Keepon is even able to dance!

Keepon started to dance during the summer of 2006 when Marek a PhD student at Carnegie Mellon interned under Kozima. Kozima placed Keepon on Marek's desk and asked him what they could do with it? Marek replied, " Let's make it Dance!" Marek then made Keepon to dance to the tune of "I Turn My Camera On" by Spoon.

What originally started as a video made just to show friends back home of what he did over the summer, Marek's Keepon video soon caught the media's attention. Wired Magazine and Spoon flew down to Tokyo for the filming of the song "Don't You Evah" from their new album "GaGaGaGaGa." Keepon is shown dancing around Tokyo with his developer Hideki Kozima. Then, in Denmark, Michalowski attended the Robots at Play festival, exhibiting Keepon & the BeatBots—a robotic dance troupe which move in synchrony with music using Nintendo Wii controllers. For this event Keepon won a 10,000 Euro Prize and in Korea, Michalowski won Best Interactive Demonstration at the RO-MAN 2007 conference. Keepon  danced at the Spoon concert in Los Angeles and was also showcased at the WIRED NextFest this year.

To see Keepon in action, watch it dance to "I Turn My Camera On" (http://www.youtube.com/watch?v=3g-yrjh58ms) and "Don't You Evah" (http://www.youtube.com/watch?v=nPdP1jBfxzo)

 talkback to the pulse