RoboSegway to the Rescue!
Computer Scientists Help Build Robotic Segway
It's almost as tall as an average person, and its head is a single
"eye" in the form of a digital camera. Its heart is an open book-a
laptop with the screen up-and it moves around autonomously on a base
with two sturdy tires.
It's a robotic Segway, and right now it can follow cars, strolling
people, and pets. But faculty and student researchers in the UMass
Amherst Laboratory for Perceptual Robotics (LPR) are creating control
software that they hope will one day enable the free-wheeling Segway to
help transport injured persons out of disaster sites and assist
astronauts in planetary exploration.
The robotic Segway is an autonomous version of the Segway Human
Transporter (HT) called the Segway Robotic Mobility Platform (RMP). The
Segway HT is the two-wheeled platform on which a person stands and
moves up to 12 miles per hour by shifting his or her weight. Segway has
collaborated with the Defense Advanced Research Projects Agency (DARPA)
to create an autonomous platform that can be driven by a computer,
instead of a human. On the platform are sensors and computers that the
Segway RMP automatically balances using rate gyros to measure the
attitude of the platform.
The UMass Amherst LPR research on the Segway RMP is funded by a grant
to Computer Science Professor Rod Grupen from DARPA, and the Segway RMP
is on loan to the department for a year for testing and development of
advanced control software. The UMass Amherst team is one of a dozen in
the country working with the modified Segways to develop a
high-performance, human-scale mobile platform that can collaborate with
humans to reduce their exposure to hazardous environments, follow them
and carry heavy loads, or carry sensors that boost a human's
perception. The platform currently has a camera and laser range finder
mounted on the Segway to follow a human leader while avoiding
obstacles. Student and faculty researchers have demonstrated outdoor
follow-the-leader tasks at speeds previously unattainable with
commercially available mobile robots.
Oliver Brock, assistant professor of computer science and a member of
the research team, says, "Our Segway is capable of a number of skills
necessary in order to transport injured people autonomously: avoiding
obstacles, detecting impassable terrain, finding new routes, and
following leaders."
Andrew Fagg, senior research scientist in computer science and another
member of the team, says the goal is to develop a versatile autonomous
platform that can be used in a variety of applications. "We are working
to extend these techniques to allow the robots to automatically perform
search-and-rescue tasks with limited human interaction. This will allow
many robots to be used to find survivors in damaged buildings, giving
human rescuers a good sense of where help is needed and where the
hazards are even before they enter."
The UMass Amherst team hopes ultimately to mount arms and hands on the
platform to allow the robot to perform tasks shoulder-to-shoulder with
other robots and humans.
more:
UMass Amherst LPR Segway
UMass Research Infrastructure:
Sensorimotor Development in Humans and Machines
story by Deborah Klenotic
March 26, 2004.
