Volume 9, Number 1, Spring 1997
Topic for this issue: AUTOMATION
Wearable computer promises to revolutionize
poultry plant communications Future technologies will enhance poultry operations
Designing robots for the poultry industry: ATRP researchers turn to academic
experts for help
Military technology adapted for visual product inspections
Advanced sensors improve efficience
1997 Poultry Show a success!
Wearable computer promises to revolutionize poultry plant communications
Georgia
Tech researchers have developed a wearable computer system for the
poultry industry. The wearable computer is similar to a standard desktop
computer, but in miniature size. It allows plant workers to enter
and
receive information from anywhere in the plant. Georgia Tech researchers
are currently field testing the system and planning for future improvements.
A
quality assurance worker on a poultry plant floor inserts a probe into
a passing chicken breast to measure its internal temperature,
but instead
of taking the time to pick up a clipboard and write down the information,
he simply says the temperature aloud. A tiny microphone on the worker's
headset (similar to a telephone operator's) transmits his voice data
to voice-recognition software on a small computer that the worker
comfortably wears on his hips. This wearable computer records the
temperature into
a
computer database file and instantly transmits the file over a wireless
communications link to quality control's computer. The quality assurance
manager taps at her computer keyboard and the product temperature
data she needs to complete an important quality assurance report instantly
appears
on her computer screen. Does this sound far-fetched? Georgia Tech
researchers
don't think so. They have developed a customized wearable computer
system for the poultry industry that does just this.
Poultry processing
today requires powerful knowledge bases and timely information
to maintain efficient operations. Floor managers often
need immediate access to information from other areas of the plant
to make more-informed
decisions, yet that information may not be available when and where
they need it. Wearable computer technology allows a worker on the
plant floor
to enter and receive information in real time (real time means making
information available at the time and place it is needed). In doing
so, the system also
provides plant management with the latest operations data about what
is going on in the plant. This technology promises to be a powerful
tool in
keeping a plant running smoothly in this era of Hazard Analysis Critical
Control Point programs and statistical quality control.
Computer design The Georgia Tech design team includes an industrial engineer,
a human factors engineer, and computer specialists who designed the
computer system with user comfort, ease of use, and low cost as priorities.
Drawing from
each of their fields of expertise, the team members integrated
existing
hardware and software technologies to develop the wearable computer.
The
wearable computer consists of all the major components of a standard
desktop computer, but in miniature size. A user wears the
computer
(about the size of a portable cassette player) on a padded adjustable
belt. A battery
pack to power the computer also fits on the belt and can be bent
slightly to fit comfortably on the user's hips. The battery lasts
four to six
hours and can easily be replaced while the computer is running.
A user also wears
a headset consisting of a microphone for voice entry, an earphone
for audio feedback, and a tiny computer screen for visual interaction.
Combined
with
customized applications and a wireless communication system,
the computer allows a user to enter and receive hands-free information
from virtually
anywhere in the plant.
Voice entry of data was a key feature deemed
important to system success. Yet plant noise levels in operations areas
presented
serious challenges
to voice recognition. "We had to evaluate a number of commercial
speech recognition systems to find one that worked well in
such an environment," says
Chris Thompson, Georgia Tech research engineer. The team tried
to replicate the particular noises of a poultry plant but found
this difficult in
a laboratory environment, so they relied more heavily on their
tests in the plant.
As user comfort is a big concern for the
design team, balancing the tradeoff between battery weight
and performance is important
to the
design process.
The team now uses a nickel metal hydride battery that lasts
approximately five hours and can be comfortably worn for a
length of time by
a user. Lithium ion batteries are available that give higher
energy per
weight
and can last
five hours longer than the nickel metal hydride battery. A
major drawback, however, is that if mishandled, lithium ion
batteries can catch on
fire. The team is continuing to look for more efficient and
safe
batteries.
Initial field trials underway
"
The wearable computer system could transform how daily operations take
place in a poultry plant," says Thompson. "We are currently field
testing the prototype at Cagle's plant in Camilla, Georgia to see how
it performs on quality control data entry."
"
It is very exciting what this type of system could do for our quality
department as well as many other areas of the plant," notes
Milton Lowry, Quality Manager at Cagle's plant. The design
team has generated
several of the plant's quality control forms in electronic
format. This has required
some ingenuity by the designers because the information must
be brief and readable on the small computer display screen
yet connected just as
it is
on the printed form.
Thus far, field tests for
recording temperature data from products into the wearable
computer have gone smoothly, and
Thompson
anticipates this
will be an attractive use of the wearable computer in poultry
plants. In the next phase of field testing, Thompson's
team will evaluate
how easily
poultry workers are able to use the computer system. Cagle's
employees will use the computer system in their plant operations
and provide
feedback to
the Georgia Tech team.
Future Improvements
As computer hardware continues to get smaller and lighter,
Thompson says the wearable computer may one day fit
in a shirt pocket.
Voice recognition technology also continues to improve,
so the wearable computer will
be able
to more clearly identify what a worker is saying in
a noisy environment such as a poultry plant. The tiny monitor is
currently black
and
white, but the wearable computer will incorporate a
color headmount display
as these decrease in price and improve in performance;
color displays currently
cost three times as much as monochrome displays and
are typically lower in resolution. Batteries also continue
to improve and
will likely weigh
less and last longer in the future.
Thompson and his
team are also looking at other potential enhancements to the system
that include adding video
capability and supporting
the maintenance function in plants by providing not
only databases at the
plant level, but
also longdistance real-time linkages with equipment
manufacturers and experts anywhere in the world.
Beyond the poultry industry, the National Aeronautics
and Space Administration (NASA) has expressed interest
in using
the wearable
computer for
technicians who perform maintenance on the space
shuttle. The airline industry
may also be interested in it for technicians who
repair aircrafts.
In the future, wearable computers
may routinely be used to provide multimedia information in the
form of text,
graphics, video,
sound, and animation to
mobile workers as they actually perform a task.
Other applications of the technology will likely develop
as poultry processing
plants become more
automated.
Future technologies will enhance poultry operations
In the poultry processing
plant of the future, machines will communicate with each other and
automatically adjust their own controls for better performance.
Human workers will track quality control automatically using data transmitted
from machines and plant floor workers and, with a few keystrokes, control
plant operations. Such futuristic plants may be closer to reality than
you think. Advances in automation technology promise to drastically
change how
poultry plants operate in the coming years.
Computers will likely allow
control and monitoring of equipment operations automatically, with
human workers overseeing these operations and supervising
control strategies. The days when human workers have to manually
change machine settings on processing equipment may soon be a thing of
the past. For example, when a plant cooks a product, different pieces
of
cooking equipment
will be able to communicate with each other about the condition of
the product. If sensors indicate a ready-to-eat product needs more
salt, a transmitter
will send a radio frequency signal to a device that will automatically
add more salt to the product. If a temperature-reading device finds
that cooking
operations are not hot enough, the device will signal an oven or
boiler to increase its temperature. These automated interactions between
pieces
of equipment will increase the productivity of the plant and will
also allow more time for the human work force to check quality and maintain
operations
for optimal performance.
Although maintenance of such equipment will
pose greater challenges, it will also become more automated and easier
for workers to perform
because processing equipment in the future may have a computer
screen built right
into the piece of equipment itself. This screen will display
information about the piece of equipment such as repair manuals, possible
causes
of maintenance problems, or step-by-step multimedia (text, drawings,
animations, video, and sound) instructions for repair tasks that
will aid workers
in
its maintenance. Other smart technologies built into equipment
will also help workers in daily operations.
In another use of the latest
electronic technologies, plants are already beginning to take advantage
of electronic information
both with one
another and with suppliers and distributors. This activity
can be expected to expand
greatly. For example, information about a growout flock about
to be delivered to a processing plant could be automatically transmitted
the
day before
the flock is picked up so that the processing plant has a profile
on average weight, health, potential risk of pathogen contamination,
and
the like before
the flock arrives. This would better prepare the plant to handle
that flock.
Much of this technology is either emerging or already
available. Poultry plants must focus on finding ways to economically
incorporate them
into daily operations. As the industry becomes more sophisticated,
daily operations
should become less labor intensive, with technology improving
efficiency, productivity, and competitiveness.
Designing
robots for the poultry industry: ATRP researchers turn to academic
experts for help
In their development of a human-level performance robot
for the poultry industry, ATRP researcher Gary McMurray and his research
team (which
includes several graduate students and professional engineers) have
called on the
expertise of academic professors Kok-Meng Lee and Harvey Lipkin -
recognized experts in the fields of robotics and dynamics - to help them
develop
new ways of designing and controlling robots. This collaboration
between ATRP
researchers with their practical expertise and academic experts with
their more specialized expertise is helping ATRP researchers tackle
more effectively
the challenges that affect the ability of robotics technology to
be used on a processing plant floor.
A specially designed robot for the
poultry industry
For the past several years researchers in Georgia Tech's
Agricultural Technology Research Program have been working to advance
robotic technology for the poultry-processing floor. Some industries
have benefited from
robots on the manufacturing floor for some time. Industries that
have been successful
in using robots are those that make parts that are consistently
shaped and easily handled - such as automobile parts and electronic
printed
circuit boards. The poultry industry, however, has parts that
come
in
all shapes
and sizes, that are soft and deformable (and hence, not easily
handled), and that can be damaged if not handled appropriately.
A
robot for the poultry floor, then, must be able to
-
identify what type
of part it must pick up;
-
identify where that part is located;
-
grasp products of varying sizes;
-
change its gripping strength so that
it doesn't damage the part;
-
withstand the environment of the processing
plant, which includes washdowns.
Traditional robots designed for other industries just haven't met these
requirements at a price that would make their introduction onto the poultry
floor economically justifiable.
According to ATRP researcher Gary McMurray,
human-level performance robots are a possible solution. Human-level
performance robots are designed to
approach tasks with human-like capabilities. For example, humans are
smart: they don't have to be programmed (told how to do each task)
as a robot does
because humans can learn. Furthermore, humans are extraordinarily successful
at pick-and-place tasks: humans can see where objects are and can adjust
their motions to pick up a moving object, like a poultry part on a
conveyor belt. And humans are adaptable: their body designs let them
perform many
different types of tasks.
The ATRP research teams have set out to design
robots that would incorporate many of the features that we humans take
for granted in performing
varied tasks - intelligence, adaptability, and vision. The research
teams are currently
working on two prototypes - the IIBM (Intelligent Integrated Belt
Manipulator) and a visual servo unit. Developing such robots has recently
brought
together researchers in GTRI's ATRP program and Georgia Tech academic
experts in
the fields of robotics and dynamics. Together, they have forged strong
collaborations to meet some of the challenges that developing human-level-
performance
robots presents.
The IIBM
One prototype robot - the IIBM - that McMurray and his ATRP team
have developed over the past two years includes many human-level-performance
attributes. For example, the IIBM has an end effector - its hand
-
that has suction cups for picking up tray packs. This end effector
can change
its shape and the pressure it exerts to compensate for varying
sizes of tray packs, much as a human hand would. A tracking system
is used
to provide "eyes" for
the robot, allowing it to identify different types of products
and their location on the conveyor. This information is then transferred
to the robot
to allow it to move to the correct position to manipulate the part.
The
robot arm, to which the effector is attached, can move in three
directions. It can move along a horizontal beam 60 inches long;
it rotates plus or minus
165 degrees, allowing the robot (after the end effector has grasped
the pack) to orient the tray pack inside the box; and the end effector
can move
up or down 5.25 inches. When the IIBM is placed over a conveyor
belt, the robot's arm can move along the beam to position itself
over a
tray pack,
rotate to be in the proper position to pick up the tray pack, and
lower the end effector to pick up the tray pack. The robot then
places the
tray pack in its proper position and orientation in the shipping
case.
In their previous tests of the system, McMurray and his coworkers
had achieved an 0.6-second move along the horizontal beam and a
1.5-second-per-pack cycle time, with 22 inches of movement in the
horizontal, 5 inches
in the vertical, and 90 degrees of rotation. The researchers felt
that
the robot
had the potential to perform tray-pack operations under plant conditions.
McMurray notes, however, "The system vibrated under certain conditions,
and we felt we needed higher speeds. We needed to develop some
new-and-improved control algorithms, and we needed an expert in
this area. Dr. Kok-Meng Lee
with Georgia Tech's School of Mechanical Engineering had the expertise
we were looking for, and he had ideas on how to solve the issues." (Algorithms
are the equations that produce the control signals to move the
robot, based on the data received from the robot's sensors.)
Solving
problems with the IIBM
McMurray was concerned about the robot's
tendency to vibrate if it moved beyond a certain speed. Dr. Lee involved
one of his graduate
assistants, Chad Rutherford, to do the modeling on the IIBM robot's
horizontal
motion
under Dr. Lee's guidance. This is the most time-critical motion
and
the most complicated drive system of the robot's movements. Rutherford
worked
with what is called a frequency-reshaping algorithm, which achieved
two important goals, according to McMurray: first, it improved
the cycle time
of the robot, decreasing the standard horizontal motion to under
0.5 seconds; second, it was able to make the robot less likely
to vibrate.
To achieve these goals, Rutherford needed to model
the physical parameters of the robot. For example, the robot does not
uniformly accelerate
and decelerate - it starts off slowly, rapidly accelerates,
and then decelerates before
picking up the package. The model needed to incorporate the
friction of the parts against each other, the motor design, and the way
the system itself
is built - all of which affect how the robot performs. Rutherford,
under Dr. Lee's guidance, was able to create an experimental
model over a small
range of inputs that took into account all the physical parameters. "This
model, which was accurate over a small range, served as a benchmark
for the more complete theoretical model that Rutherford developed
as part of
his master's thesis," McMurray emphasizes.
Next, Dr. Lee and
the research team, of which Rutherford was now a part, constructed
a complete theoretical model for all
the elements
of
the drive
system. The model of the elements was based on the manufacturer's
data for the motors, gear boxes, and the oil- filled bearings,
for example. However,
McMurray explains that many of the elements of the design had
to be
modeled by the team. He sighs, "There was a lot of tedious
modeling of the elements, such as the damping effect of the
oil-filled bearings." (Damping
is the resistance that the part encounters during operation.) The
research team put all this information together using a mathematical
package called MATLAB™. McMurray and the team were delighted
with the results: "We were able to model the robot system
in complete detail and to get the complete mathematical model
to agree with our experimental
data, which is a major triumph. We can actually model the robot
under any circumstances that we want and start changing components
in the model. For
example, if I want to change out the gear box, I can put a
new gear box in the model and see how it would affect the system
almost as if I were
physically running the system. We can make all kinds of changes
to optimize the system without the tedious, time-consuming,
and limited experimental
trials." What the model also allows the researchers to do
is to try out the control algorithms - the various control
ideas - in real time and
see how the robot would perform. Through modeling, the research
team can simulate a wide range of operating conditions, some
that could potentially
damage the robot or cause excessive wear and tear under real-world
test conditions.
Improving performance on the visual servo
robot
McMurray and his team not only were interested in solving
the issues associated with the IIBM but also had begun another
project to
improve the sensor inputs of robots (their "eyes")
in general. McMurray and his coworkers wanted to mimic the sort
of
hand-eye coordination that a human
uses in performing tasks.
McMurray notes how important this
sort of research is to overcoming some of the limitations
of placing robots in food
processing
plants: "Robots
are traditionally blind machines that are preprogrammed to
move from point 'a' to point 'b' in a very fast, accurate
manner. If the part is outside
of its preprogrammed position, the robot assumes that the
part isn't there, and the robot doesn't know what to do." McMurray
cautions that food products can vary in position, orientation,
size, and sometimes product
type: in a poultry processing plant, for example, product
can come down mixed - wings and thighs. To cope with what
would be a perplexing situation
for a robot, a human just looks down at the part and picks
it up in the appropriate manner.
McMurray and his research
team are developing a system so that the robot can "see" the
way a human does and respond accordingly - it could adjust
its movements based on what it "sees." The human looks
at his/her hands, looks at the target (the piece of poultry),
and moves based on what he/she sees. The person makes small
corrections as he/she
moves and as the part moves. Unfortunately, current robots
can't match this level of flexibility and adaptability .
To
try to incorporate these attributes into robots, McMurray
turned to Dr. Harvey Lipkin with Georgia Tech's School of
Mechanical Engineering
for
his controls expertise and background in kinematics (the
study of the motion of mechanisms). As a graduate student,
McMurray had worked
with
Lipkin,
and they had maintained interest in each other's research
projects
over the years. Lipkin agreed to serve as a mentor to two
graduate students, Jenelle Piepmeier and Rob Biro, who were
part of
the ATRP research
team.
Their task was to design the control algorithms and to build
a small concept robot that would be controlled by these new
algorithms.
McMurray stresses the importance of the research: "If
we can develop this visual servo robot and control algorithm,
we can start building low-cost
robots because we don't need the incredibly accurate joint
sensors and the rigid construction that drive up the cost
of current robot technology." He
states, "What we are trying to build is a low-cost, simple,
slow robot that will have varying joint lengths and inexpensive,
low-accuracy sensors.
We will be demonstrating the robustness of the algorithm
under the varying kinematic changes in this type of robot."
Biro,
under Lipkin's guidance, is currently designing the prototype.
His design consists of a horizontal robot with
three degrees
of freedom (three movements), all rotation axes. It is a
very simple mechanism
that can actually be adjusted to have various joint lengths.
This
type of adjustment
would be analogous to a human's being able to change the
length of the arm from the elbow to the wrist to pick up
a pencil, McMurray explains. Biro
plans to test the varying lengths of the robot's "arm" to
see how the robot manages the changes.
Piepmeier, also under
Lipkin's guidance, is working on the algorithms that can
control such a visual servo robot. She
is investigating
simulation work performed by previous researchers, identifying
(with Lipkin's
assistance) limitations in the previous research and current
theory, and making suggestions
for improvements.
McMurray compares the theory with which
they are dealing to the complexity of downhill skiing: "Imagine
that you are at the top of a very difficult ski slope and are trying
to get down to the base. You can't map the entire
way down the ski slope because you can't see the entire terrain.
As you start to move, your algorithm needs to be able to guide
you down that path
(and keep the base in sight) without misleading you down
false paths. The current algorithms we are using are one-dimensional,
but our ski problem
is a three- dimensional problem. Our one-dimensional algorithms
work well in a limited small area; but if we have a big area like
our ski slope, then
the algorithm begins to fail. For example, imagine the ski
slope with a sudden drop that you can't see until you've gone over
it - we have a problem
like that with our current algorithm: we call it a singularity
position. So we have to keep the robot's motions very small and
examine very small
portions at a time."
McMurray, coordinator of the project,
and his academic collaborators hope to begin constructing
the prototype this summer and
start testing some of the controls in the fall. The visual
servo robot could add
the versatility
of the human to the robot through incorporating a type of
human
vision. McMurray is enthusiastic about the visual servo design: "This
type of robot, with its ability to see and handle products
in an unstructured
environment, combined with its potential low cost, could
have a major impact on the poultry-processing floor."
Military
technology adapted for visual product inspections
Georgia
Tech researchers are using computer simulations to mimic the way
human visual processes determine if a poultry
product is defective.
The research is based on a model of human visual processes developed
by a team of GTRI research scientists to help the U.S. Army
and Air Force identify
tanks on rural and urban landscapes. This technology has strong potential
benefits for the poultry industry, since it provides an opportunity to
speed up the inspection process while maintaining a higher
degree of consistency
and accuracy.
Georgia Tech researchers are adapting computer technology
developed to detect helicopters and tanks that are located in the
midst of clutter
to the quality-control needs of the poultry- and food-processing industries.
The research is still exploratory, but scientists and engineers at the
Georgia Tech Research Institute (GTRI) have made progress in getting
a computer simulation of a human quality inspector to distinguish between
defective
and first-quality poultry products and fruit products.
The computer simulation
used in this project mimics the way human visual processes pick out
targets against cluttered backgrounds. It reviews
a series of images of food and poultry products photographed by a camera.
With the
help of previously stored calculations concerning color, shape, and
texture of poultry or fruit, the computer determines if the product
is defective.
The simulation is based on a model of human visual
processes developed by a team of GTRI research scientists, headed
by Dr. Ted Doll, to help
the U. S. Army and Air Force identify tanks on rural and urban landscapes.
The
team has been looking for ways to apply the Georgia Tech Vision Model
(GTV) to an industrial processing operation. Product inspection is
an ideal use
of the technology because this task involves "looking" for defects
just as human workers do. Computer automation has strong potential
benefits for the poultry- and other food-processing industries, since
it provides
an opportunity to speed up the inspection process while also maintaining
a higher degree of consistency and accuracy.
The Georgia Tech Vision
team began adapting their model to the needs of the poultry industry
when approached by Wayne Daley, a research
leader in Georgia Tech's Agricultural Technology Research Program
(ATRP). Since
1988, Daley has been working with computer vision on a variety of
applications for the poultry industry, including such tasks as
sizing, part sorting
and quality control.
The simulation software can pick out defects,
such as tears in poultry and scarring in grapefruit. But to be
practical in a factory setting,
it must make these identifications at a much higher speed than is
now possible. Capable human workers can inspect 35 chickens a minute
-
or, one every
two
seconds. The computer doesn't come close to meeting these speeds
yet, although Daley is presently working to increase the simulation's
inspection
speed.
Doll is confident that the simulation can achieve the required inspection
speed if it is implemented on a very fast computer or in special-purpose
hardware.
Once the simulation is operating in real time (real time
means making information available at the time and place it is
needed), it will
provide a general-purpose tool that can be adapted to any inspection
task that
a human can do. Because it imitates human vision, the vision simulation
will
be much more powerful and flexible than existing automatic inspection
tools based on artificial neural networks. Based on the technology's
success in
the past and the promising results of the system’s adaptation,
researchers are optimistic about the future benefits to the poultry-
and food-processing
industries.
Advanced sensors improve efficiency
Georgia Tech researchers are developing
smart sensors to assist the poultry industry with the issues of parts
tracking and quality inspection. The parts-tracking
system records the parameters of a product and tracks it in time on
a conveyor belt. The integrated color machine-vision system examines
a product
for
visual defects to determine quality. Both machines are designed to
perform functions that currently require human vision.
Just as humans
use their eyesight to locate an object when moving it from one
place to another, a machine also requires sensors to be as
accurate as possible when moving products on a conveyor belt or when
checking products
for irregularities. For the food processing industry, these sensory
machines can be useful when determining meat quality and when handling
materials.
Today, these tasks frequently must be performed by humans, but the
job is often tedious and boring. In an effort to assist the food processing
industry
in this area, researchers with Georgia Tech's Agricultural Technology
Research Program have developed a parts-tracking system and are currently
developing
an integrated color machine-vision system. These smart sensor systems
will assist companies by tracking the position of products on an assembly
line
and by examining the appearance of a product to determine product quality.
"
We are interested in helping the industry improve its overall efficiency
in a cost-effective manner," says Wayne Daley, Georgia Tech senior
research engineer. "These devices are designed to help machinery
function in intelligent ways at reasonable costs."
The two systems
operate by using devices such as photodiodes and CCD (Charge-Coupled
Devices) arrays, which detect shape and quality
details
much as the human eye does. The equipment has an advantage over
the naked eye however, in that it also employs computers that do
not
tire after hours
of performing repetitive tasks such as inspecting meat for irregularities
or checking the position of products on a line. Also, in the poultry
industry's factory environment, which can often be harsh due to
water and cleaning
agents, the size and durability of these compact devices can be
important.
The parts-tracking system is a low-cost alternative to
computer vision systems for locating and identifying products
on a conveyor
belt.
It uses an array of photosensors to determine the position in
time of a product and also records the shape and orientation of
the
product. It
can then
supply
this information to a handling machine, which can use the data
to
properly and safely move the product from one location to another.
The parts-tracking
system can also identify a problem on a conveyer line. For instance,
the system can recognize if a product is missing or not properly
positioned.
The integrated color machine-vision system is designed
to be a cost-effective, self-contained imaging system. Prior
to its development,
the technology
used three separate and bulky systems. First, a camera took
a black
and white picture of the product. A frame grabber then digitized
the picture to put it into the memory of a computer, which
analyzed the picture.
With
the integrated color machine-vision system, the process can
now be done in one step and has the added benefit of being in color.
Because of this
new technology, if any features of the product are unsatisfactory,
such as miscolored or bruised skin on a piece of meat, the
machine
is
quickly
able to identify the product quality as undesirable.
Georgia
Tech is developing these systems in collaboration with local
industries. The parts-tracking system, which the researchers
are
in the process of patenting, is now being evaluated by a
northeastern robot company
for commercial release. The integrated color machine-vision
system is being produced in association with an Atlanta-based
computer
vision company and
a Georgia-based food equipment manufacturer.
Currently, the
parts-tracking system is in use in Georgia Tech's own robotics
research for the Georgia Poultry Industry.
Daley
predicts that the integrated machine-vision system will
be ready for commercial
use in
another year or two. He anticipates that both of these
devices will prove to be extremely beneficial to the food processing
industry because
they
can increase the accuracy of process control and product
quality in the factory.
Back to Topics
1997 Poultry Show a success!
Georgia Tech Agricultural Technology Research Program scientists and
engineers met with visitors to the 1997 International Trade Show and Exposition
in Atlanta on January 22-24, 1997.
ATRP displayed several of its research projects including the wearable
computer. Visitors to Georgia Tech's booth were able to try on the wearable
computer and imagine themselves using it in a poultry plant.
We hope to see you at next year's show!
Credits
Dara O'Neil, Editor
Rae Adams, Contributing Editor and Photographer
Margie Brown, Contributing Editor
Nancy Davis, Contributing Editor
Jim Demmers, Contributing Photographer
Caroline Fitzpatrick, Contributing Editor
Mark Hodges, Contributing Editor |
