GEERHEAD
Photo from the day
that Drexel
University set up
four of the HUBO
robots to pose as if
crossing the street
like the Beatles
from their Abbey
Road album.
12 SERVO 10.2012
Some videos of HUBO on You Tube
include the robots playing “Come Together”
from the Abbey Road album by the Beatles.
Dr. Kim and colleagues took a photo of the
HUBO robots crossing the street so that it
looked like the Abbey Road album cover. The
photo appeared on the cover of the Drexel
online magazine last spring.
While Dr. Kim has developed the beat
tracking capabilities for HUBO, he has not
perfected the robot’s dancing capabilities.
“The HUBOs cannot dance in a good,
realistic way yet. They can perform
programmed gestures at the right time.
The goal is to have them perform the right
dancing gestures linked together in the right
way with some creativity,” says Dr. Kim.
The gesture and motion planning for
dancing is a very complicated problem
that is similar in complexity to the motion
planning for having robots work with tools
or operate doors and latches.
The long-term goal of Dr. Kim’s research
is to have the robots play musical
instruments alongside human musicians in
an ensemble. “We eventually want robots to
be autonomous enough to be personal
assistants, help people get around, do
dishes, clean up, and fold laundry. We are
taking steps in that direction, as well,”
comments Dr. Kim.
The HUBO robot is physically capable
of these things. It can deal with curbs and
stairs (walking), and can use its fully
actuated hands to pick up light objects,
including tools. “These are all individual
capabilities. Linking them together takes
baby steps,” expressed Dr. Kim.
Advances
Most recently, Dr. Kim and colleagues
have made a lot of strides in the listening
capabilities of the HUBO 2 robot. As the
robots move, they make a certain amount
of noise, so while they are trying to do
something such as move or interact with
other musicians, if someone is talking in the
background as well, the motor noise makes
it difficult for the robot to discern one sound
from another. “We have worked out a
couple of different algorithms to rid the
noise,” explains Dr. Kim. He is also using
better microphones that are directional, and
filter out unwanted noise while filtering in
the desired sounds. “We are working on a