Ateneo Physics student Mariel Dee programs the NAO Robot at Nara Institute of Technology, Japan

Mariel Dee and the NAO Robot in NAIST, Japan

Mariel Dee and the NAO Robot in NAIST, Japan

by Quirino Sugon Jr.

It is a creature perhaps from an another world clad in white armor like a Storm Trooper shorn of its helm, revealing two large ears like a stereo speaker but glowing like Tron disc blades.  Its deep set eyes stared into space like black coals burning in blue flame.  Its three-fingered hand grasped the air like claws, as if wielding an invisible light saber.  And from lips unseen, save for a tiny hole beneath its eyes, the creature spoke in a child-like voice reminiscent of the White Queen in Resident Evil:

“My name is NAO.”

NAO is an autonomous, programmable humanoid robot developed in 2004 by a French startup company Aldebaran Robotics.  NAO stands a little less than two feet and weighs like a baby.  With about 21 to 25 ways to move–the swaying and nodding of the head, the rotations of the arms and legs, the grasping of the fingers–NAO can walk like a toddler, pick up an object, or play football in RoboCup.

Mariel Dee, a BS in Applied Physics with Applied Computer Systems student in Ateneo de Manila University, had programmed the NAO robot during her internship at Nara Institute of Technology in Japan last 14-27 October 2012.

“One can program NAO,” she said, “by typing text commands directly to its computer processor.  But I programmed NAO using the Kinect.”

Kinect is a motion sensing input device developed by Microsoft for the Xbox 360 video game console.  Kinect has a camera that receives red, green, and blue (RGB) colors as input.  Kinect also has depth sensor which consists of an infrared laser and a CMOS APS–an active pixel sensor that determines the noise level at each pixel and uses its circuitry to cancel it out.  Kinect is the technology that powers games like Dance Central where you try to match the dance moves of your virtual dancer on the computer screen.  The closer your moves are to your virtual dancer, the higher your score.

What Mariel did is to make Kinect read her movements, translate them into skeleton frame motion, and use this as input in a language that NAO can understand.  The Kinect skeleton frame is made up of several joints.  The central spine consists of the head, shoulder center, spine, and hip center.  The arm  consists of the shoulder center, shoulder (side), elbow, wrist, and hand.  The leg consists of the hip center, hip (side),  knee, ankle, and foot.  In all there are 20 joints connected by 19 line segments.

In her program, Mariel made use of a limited number of NAO’s movements: walk forward and backward, walk sideways to the left and right, turn left and right, kick with right leg, bend down, push with hands.  But NAO is like a deaf boy trying to learn his first sign language.  For example, if  Mariel raises her right hand forward, NAO walks forward; if backward, NAO walks backward.  If Mariel raises her right hand to the sides, NAO walks sideways to the right; if her left hand, NAO walks sideways to the left.

“The main problem is how to make the program robust,” said Mariel.  ”NAO should be able to obey commands even if it were posed by child or an adult whose heights and arm lengths are different.  So what I did is to compute the angles formed at the joints and use this as a Kinect input for NAO.  In this way, the lengths such as that  between the elbow and the wrist or from the shoulder to the elbow would not matter but only the angle formed by these two segments at the elbow.”

 The skeletal frame seen by Kinect looks like molecules in Chemistry, with the joints represented by atoms and segments by bonds.   In a way, what Mariel wishes to measure are the equivalents of body angles of a molecule.  For example, the body angle for a water molecule H-O-H at the Oxygen atom is 104.5 degrees.  But unlike molecules which have more or less fixed body angles, the body angles in a NAO robot have greater range of possible angles.

So can we now have robot combats as seen in old animes like Tosho Daimos or in more recent movies like Real Steel, with robots mimicking the actions of their human controllers?

“No.  We are not yet there,” said Mariel.  “But I am glad that after I left, the NAO robot that I programmed was showcased during an open house of the NAIST Robotics Laboratory.  NAIST is nice.  It is fun working there.  They still have other robots there like a humanoid face that mimics your emotions.  But NAO is so cute.”


About ateneophysicsnews
Physics News and Features from Ateneo de Manila University

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: