Tech Briefs Magazine - July 2021 - 40

Robotics & Automation
workers, the human-robot collaboration
system can identify each worker as well as
the worker's skeleton model, which is an
abstract of the worker's body volume.
Using this information, the context-aware
robot system can recognize the worker's
pose and even predict the next pose.
These abilities provide the robot with a
context to be aware of while interacting.
The system operates with artificial in -
telligence that requires less computational
power and smaller datasets than
traditional machine learning methods.
It relies instead on a form of machine
learning called transfer learning, which
reuses knowledge developed through
training before being adapted into an
operational model.
With a current collaborative robot,
when a human approaches it, the robot
slows down and if the worker comes
close enough, the robot will stop. If the
person moves away, it resumes. The context-aware
robot system can be compared
to a self-driving car that recognizes
how long a stoplight has been red
and anticipates moving again. Instead of
braking or downshifting, it begins to
Algorithm Gives Robots a Faster Grasp
The algorithm speeds up the planning process robots use to adjust their grip on objects for
picking and sorting or tool use.
Massachusetts Institute of Technology, Cambridge
A
new algorithm significantly speeds
up the planning process required for
a robot to adjust its grasp on an object by
pushing that object against a stationary
surface. Whereas traditional algorithms
would require tens of minutes for planning
out a sequence of motions, the new
approach shaves this preplanning process
down to less than a second. This faster
planning process will enable robots, particularly
in industrial settings, to quickly
figure out how to push against, slide
along, or otherwise use features in their
environments to reposition objects in
their grasp. Such nimble manipulation is
useful for any tasks that involve picking
and sorting, and even intricate tool use.
Existing algorithms
typically
take
hours to preplan a sequence of motions
for a robotic gripper, mainly because for
every motion that it considers, the algorithm
must first calculate whether that
motion would satisfy a number of physical
laws such as Newton's laws of motion
and Coulomb's law describing frictional
forces between objects. A compact way
to solve the physics of these manipulations
in advance of deciding how the
robot's hand should move involves using
" motion cones " that are essentially visual,
cone-shaped maps of friction.
The inside of the cone depicts all the
pushing motions that could be applied to
an object in a specific location, while satisfying
the fundamental laws of physics and
enabling the robot to keep hold of the
object. The space outside of the cone represents
all the pushes that would in some
way cause an object to slip out of the
40
Cov
adjust its speed by cruising toward the
intersection, thereby sparing wear on
the brakes and transmission.
Experiments showed that with context,
a robot can operate safer and more
efficiently without slowing down production.
In one test, a robot arm's path was
blocked unexpectedly by someone's
hand. But rather than stop, the robot
adjusted - it predicted the future trajectory
of the hand and the robot moved
its arm around the hand.
For more information, contact David
Callahan at press@kth.se; +46 8 790 69 76.
The algorithm speeds up the planning process for robotic grippers to manipulate objects using the
surrounding environment. (Image courtesy of the researchers)
robot's grasp. The algorithm calculates a
motion cone for different possible configurations
among a robotic gripper, an
object that it is holding, and the environment
against which it is pushing in order
to select and sequence different feasible
pushes to reposition the object.
The researchers tested the new algorithm
on a physical setup with a three-way
interaction in which a simple robotic gripper
was holding a T-shaped block and
pushing against a vertical bar. They used
multiple starting configurations, with the
robot gripping the block at a particular
position and pushing it against the bar
www.techbriefs.com
ToC
from a certain angle. For each starting
configuration,
the algorithm instantly
generated the map of all the possible
forces that the robot could apply and the
position of the block that would result.
The algorithm's predictions reliably
matched the physical outcome in the lab,
planning out sequences of motions -
such as reorienting the block against the
bar before setting it down on a table in an
upright position - in less than a second,
compared with traditional algorithms that
take more than 500 seconds to plan out.
For more information, contact Abby
Abazorius at abbya@mit.edu; 617-253-2709.
Tech Briefs, July 2021

http://www.techbriefs.com http://www.abpi.net/ntbpdfclicks/l.php?202107TBNAV

Tech Briefs Magazine - July 2021

Table of Contents for the Digital Edition of Tech Briefs Magazine - July 2021