Concerned family members
can access the unit and visit their elderly parents from any Internet connection,
including navigating around the home and looking for Mom or Dad, who may not
hear the ringing phone or may be in need of assistance. Doctors can perform
virtual house calls, reducing the need for travel.
“For the first time,
robots are safe enough and inexpensive enough to do meaningful work in a residential
environment,” says computer scientist Rod Grupen, director of UMass Amherst’s
Laboratory for Perceptual Robotics, who developed project ASSIST with computer
scientists Allen Hanson and Edward Riseman.
The robot, called the uBOT-5,
could allow elders to live independently, and provide relief for caregivers,
the medical system and community services, which are expected to be severely
stressed by the retirement of over 77 million Americans in the next 30 years.
There is no mistaking the
uBot-5 for a person, but its design was inspired by human anatomy. An array
of sensors acts as the robots eyes and ears, allowing it to recognize human
activities, such as walking or sitting. It can also recognize an abnormal visual
event, such as a fall, and notify a remote medical caregiver. Through an interface,
the remote service provider may ask the client to speak, smile or raise both
arms, movements that the robot can demonstrate. If the person is unresponsive,
the robot can call 911, alert family and apply a digital stethoscope to a patient,
conveying information to an emergency medical technician who is en route.
The system also tracks what
isn’t human. If a delivery person leaves a package in a hallway, the sensor
array is trained to notice when a path is blocked, and the robot can move the
obstruction out of the way. It can also raise its outstretched arms, carry a
load of about 2.2 pounds and has the potential to perform household tasks that
require a fair amount of dexterity, including cleaning and grocery shopping.
The uBOT-5 carries a Web
cam, a microphone, and a touch-sensitive LCD display that acts as an interface
for communication with the outside world. “Grandma can take the robot’s
hand, lead it out into the garden and have a virtual visit with a grandchild
who is living on the opposite coast,” says Grupen, who notes that isolation
can lead to depression in the elderly.
Grupen studied developmental
neurology in his quest to create a robot that could do a variety of tasks in
different environments. The uBot-5’s arm motors are analogous to the muscles
and joints in our own arms, and it can push itself up to a vertical position
if it falls over. It has a “spinal cord” and the equivalent of an
inner ear to keep it balanced on its Segway-like wheels.
The researchers wanted to
create a personal robot that could provide many services, such as a medical
alert system, or the means to talk to loved ones, all in one human-like package,
according to Grupen. To evaluate the effectiveness of potential technologies,
the research team worked with social workers, members of the medical community
and family members of those in elder care.
The collaborative effort,
dubbed project ASSIST, involved researchers from the Smith College School for
Social Work, the Veteran’s Administration (Connecticut Health Care System,
West Haven campus) and elder care community centers in western Massachusetts.
Through focus groups, the researchers learned about the preferences of potential
users.
Graduate students Patrick
Deegan, Emily Horrell, Shichao Ou, Sharaj Sen, Brian Thibodeau, Adam Williams
and Dan Xie are also collaborators on project ASSIST.
Home
