Canadian Packaging magazine - September 1998
by T.J. Becker
Game of Chicken
"The
nice thing about this type of automation is that it can eliminate low-skill
labor activities which the industry has the greatest difficulty filling
on a day-to-day basis."
– Craig Wyvill, ATRP senior researcher
 Robots have been a boon in the
automotive and electronics industries for many years, but the food
industry, particularly poultry processing, has not yet embraced this type
of flexible
automation.
"
One reason is that robotic systems on the market aren't completely compatible
with poultry needs," says Craig Wyvill, director of Georgia Tech's
Agricultural Technology Research Program (ATRP) in Atlanta, Ga.
"
Existing robotic systems tend to be overkill. They're too complex, which
makes them expensive to purchase and expensive to maintain. Compounding
the problem, the industry needs robots that can withstand the rigors of
the food processing environment, which typically adds to their cost."
Many of the jobs in poultry processing consist of materials handling
tasks, such as moving products from a conveyor belt to a box or another
conveyor belt. These are areas where robots are ideal substitutes for
human hands. Although poultry plants are using simple forms of fixed automation,
these machines have very limited capabilities.
"
Our goal was to develop a low-cost robot that could perform materials
handling with the same speed and dexterity as a human," says Gary McMurray,
a senior researcher in the ATRP and a project director for the robotics
initiative.
Enter the Intelligent Integrated Belt Manipulator (IIBM). This robot
tackles a common food industry task by removing items from a conveyor
belt and transferring them into a packing carton for shipping.
OPERATIONAL OVERVIEW
Conceived in 1992, IIBM has gone through several redesigns and refinements
over the years. The first-generation robot was powered exclusively
by pneumatics, attractive because of its low costs and ease-of-use. "Speed was good,
but the accuracy was not up to expectations," says McMurray, noting
that the prototype fluctuated up to an inch in position when picking up
items. Although poultry processing requires less accuracy than, say, chip
insertion in an electronics plant, "an inch was still too much," McMurray
says. The robot could still pick up product, but might misplace it in the
shipping carton.
The "new and improved" IIBM is a hybrid of pneumatics and electro-servo
drives. Two pneumatic axes and two electro-servo axes allow motion in four
different directions: up and down, parallel with the conveyor belt, perpendicular
across the conveyor belt, and a 90-degree rotational pivot.
In automotive and electronics industries, parts are consistently shaped
and are thus easy for robots to handle. Yet in the poultry business,
products vary considerably in size and shape, making grasping demands
another challenge
for the IIBM. Physical dimensions of the tray pack remain constant,
but the poultry pieces inside vary the contours of the package's top by
as much
as two inches, causing weight and center of gravity to shift.
"
Therefore, the IIBM's end effector had to be constructed with some flexibility," McMurray
says. Suction cups made from bellows material compress up to three-quarters
of an inch. A spring mechanism attached to the suction cups provides another
inch of compliance, allowing the grippers to conform to different contours
of product.
GAINESVILLE FIELD TEST
After four months of lab testing, the current IIBM prototype is now
being tested on the factory floor in a ConAgra (producers of Butterball
and Country
Pride products) plant in Gainesville, Ga. Speed and accuracy will be
the main focus of the field test.
"
In the lab, we don't have 1,000 different tray packs available to run
through at a time to get a true indication of accuracy," McMurray says.
Tray packs in the lab are manually fed onto the conveyor belt, with
the same group of product being used over and over. This type of lab testing
allows researchers to check on major design elements. But it can't
take
into account variability among product or the plant environment, which
may affect the robot's performance.
Case in point: The tray packs coming out of a freezer can attract moisture
in the packing room and generate frost on the package surface--affecting
how well the robot's grippers adhere to the surface.
Early field test results have been encouraging. In lab trials, the
IIBM's average cycle time was clocked at 2.1 seconds--comparable with
a human worker--and
the research team has been able to sustain this time in the plant.
More importantly, the pick-up rate of the robot has improved significantly.
During lab testing, the robot occasionally would drop a tray pack,
but missed pickup has been almost non-existent in the plant.
COSTS AND BENEFITS
McMurray estimates that final commercial costs of the IIBM will range
between US$30,000 and US$40,000--about half the price of existing industrial
robotic systems. Also, the IIBM is attractive because it is simple,
both to install and maintain. The robot can be operating after supervisors
program
only a few physical dimensions, such as the size of tray packs, the
location of packing cartons and the height of the conveyor belt.
In contrast, traditional robotic systems require as many as 300 positions
to be programmed individually. "Most food processing companies don't
have the technical base to support that type of machine," McMurray
says.
Wyvill hopes to have the technology commercialized within the next
two years. And this prototype is merely the beginning, McMurray says.
Next,
he will enhance the system with a vision system. This vision system
would help develop hand/eye coordination for the robot and allow it to
operate
by merely seeing a picture of the product, eliminating the need for
task-specific software and programming.
Does this mean a triumph of machine over man? Hardly, says Wyvill,
stressing that robotics is not about making humans obsolete. Indeed, flexible
automation
creates a better workplace for humans by upgrading job skill requirements
to a higher level. "The nice thing about this type of automation is
that it can eliminate low-skill labor activities which the industry has
the greatest difficulty filling on a day-to-day basis," McMurray says.
"
In many of these materials handling jobs, people have become extensions
of machines, which is not a good situation for the worker in most cases," Wyvill
adds.
"
Robotics make sense in these situations, freeing up a limited labor pool
and allowing humans to do what they do best--think."
SAFETY ISSUES
There's also the issue of safety concerns. Many of these jobs are highly
repetitive, putting workers at risk for cumulative trauma disorders
such as carpal tunnel syndrome. Automation can reduce injuries.
These and other benefits of materials handling robots are needed throughout
the food industry, Wyvill says.
"
We're already focusing on other ways of bringing robotics to a level
that can be widely applied across food industry lines. It will open up a
whole new world of opportunity for food equipment manufacturers and the
robotics industry."
This article was reprinted with the permission of Research Horizons--the
quarterly publication of Georgia Institute of Technology--where it
appeared in the Spring 1998 issue.
|
