Implementation of a Cartesian Robot for Remote

Mission Critical Operator Training

Aidan F. Browne, Wesley B. Williams, Keith Loftus and Cameron Nye
William Lee College of Engineering
University of North Carolina at Charlotte
Charlotte, North Carolina 28223
{aidanbrowne, wbwillia, kloftus2, cnye} @uncc.edu

Abstract— As part of ongoing work to create training for

Mission Critical Operators, a new training system was developed
offering multiple degrees of freedom (DOF), which was a desirable
addition as part of a progression of training systems of increasing
complexity. With the addition of multiple DOF, the user may be
exposed to a more complex control environment presenting
challenging tasks related to full range of motion on multiple axes.
The electromechanical training system was developed following a
previously developed standard allowing control by two separate
controllers; a Micro800 series programmable logic controller
(PLC) or a National Instruments myRIO. This gives the user an
opportunity to learn two programming languages: ladder logic
and LabVIEW. The training system includes 4 DOF and has free
range of motion in the X, Y, and Z axis. This allows for multiple
applications to be simulated. The overall approach is presented;
the final designs are being made available under an open-source
license.
Keywords—Mission Critical Operations, robot, PLC, myRIO
I. INTRODUCTION
Work has been ongoing to develop methods of training
Mission Critical Operators to troubleshoot and repair remote
equipment and systems [1]. A key element of this training is the
development of remote laboratory experiences, which have been
researched as an element of distance education since the mid-
1990s [2]. While there are many examples of remote lab
experiences that have been developed for either LabVIEW [3]
or PLC [4] programming, a goal of this project was to have a
single set of hardware that could selectively be controlled by
either LabVIEW or PLC, as indicated in Fig. 1. Previous work
includes the design and fabrication of a two car, three floor
elevator with multiple actuators, sensors, inputs, and outputs.
Control of the system uses a supervisory National Instruments
myRIO microprocessor that remains transparent to the user
operating the system. The user’s interaction with the system is
through either a second myRIO, or a programmable logic
controller (PLC) [5]. The user accesses the elevator system by
remotely and securely logging onto a virtual machine on a
dedicated server to gain direct control over the myRIO or PLC.
In addition, the user is given access to web cameras to view the
equipment while they are controlling it.
The elevator training system proved effective as an
introductory training system, as it had limited degrees of
freedom, and thus low risk for the trainee. It also established a
baseline for remote user control of electromechanical systems.
Fig. 1. Architecture for PLC and microcontroller control
Here we report on the design of another electromechanical
training system, a four degree of freedom (4DOF) Cartesian
robot, which is a step up in complexity since it allows the user
to make operational mistakes that are not recoverable (except
that the entire system is automatically resettable to a known
state). In order to teach mechatronic control as a progression, it
is desirable to gradually add degrees of freedom to the user’s
control as an individual advances through a Mission Critical
Operator curriculum. This Cartesian robot gantry allows
movement in the X, Y, and Z axis over a flat surface and a trough
of small balls, as well as control of an end effector to pick up
and place the balls. The design is similar to what might be found
in a small CNC mill or open 3D printer. The architecture of the
new system is based on the control design of the previously
reported elevator system. The baseline operational concept of
direct user control by either myRIO or PLC with supervision
from another myRIO was maintained from the previous design,
including the virtual machine access and remote web camera
viewing.
II. RESETTABLE EXPERIMENT
One requirement and challenge for developing remote
electromechanical training systems is that they need to be
resettable to a known state to allow for error recovery, as well as
to present a uniform challenge to different users. This is
somewhat trivial in a system like the elevator training system
where all moving parts are constrained in two dimensions: the

978-1-5090-2246-5/16/$31.00 ©2016 IEEE

elevator cars travel vertically on closed-ended tracks and the
elevator doors travel horizontally in closed-ended tracks.
Resetting them is simply a matter of having the supervisory
control move the motors until the cars return to a desired
position, then set each door piston to the desired state, and
finally set all output indicators to a known state. With the
Cartesian robot, the user has been afforded the ability to move
within a volume to accomplish more complex tasks than the
elevator training system allowed. This creates an environment
where the system can be commanded to move to positions or
along paths that can be destructive to the task that the system is
being used to accomplish. For instance, if the system is being
used to stack objects, then an incorrect move could cause the
entire stack to be knocked over by the end effector.
Consequently, a sub-assembly for the Cartesian robot had to
be designed that allowed utilization of the added degrees of
freedom, but was also able to be reset automatically. After much
brainstorming, many intriguing control setups were conceived,
but most had to be subsequently eliminated due to complexity
of resetting to a known state (an example of one of the desirable
concepts was a chess board system that was able to move the
chess pieces appropriately, but the task of resetting them to a
known state if knocked about was deemed to be too difficult).
The concept that was ultimately pursued was a platform that Fig. 2. Overall system
allowed arrangement of colored golf balls on its surface, as
shown in Fig. 2. The platform has the ability to do an automatic surface, regardless of color, and always guaranteeing the user a
reset by tilting to collect the golf balls at one end (using gravity), ball positioned at a known index. To accomplish this, the
and an auger design to drive all of the golf balls to known supervisory control continually uses the augers to always keep a
locations along the periphery of the table. Once reset, the XYZ ball positioned at the end of the auger as balls are plucked from
gantry, using an appropriately sized and shaped that end position. Mode Two requires the user to keep track of
electromechanical gripper, provides the means to pick up, move balls they have plucked, as the augers stay stationary after initial
and place the golf balls within the volume. When a reset is reset. Mode Three builds on Mode Two by adding the
needed, the table is tilted up or down via an electrical actuator complexity of random positioning by requiring a specific
under command of the supervisory controller; similarly the pattern of balls, inclusive of color. This requires the user to use
augers are turned with dc motors under the same control. an overhead camera and vision processing to determine the
The overall design, which is pictured in Fig. 3, proved to be position in the trough of specifically colored balls.
versatile as it can be used in multiple different modes, adding With the Cartesian robot, the user has the opportunity to
varying levels of complexity within the system. Mode One is make mistakes in programming that disarrange golf balls placed
the simplest, requiring the creation of a pattern of balls on the previously, but the user also has the ability to reset the platform
if such mistakes were made. The system also has the ability to
return itself to a known pre-set state when a user logs out, thus
presenting the next user with a known starting state
In order to ensure safety to the Cartesian robot, limit
switches were implemented to create maximum boundary limits
for each of the three axes. In order to achieve this, there are two
sets of switches: optical photo sensors and mechanical limit
switches. The reason for having two different types is to have
one set that is transparent to the user and one set that can be
controlled directly by the user. The optical limit switches are
positioned approximately an inch from the ends of travel on each
axis so that the user can learn to program them while
experiencing complexity of over-travel. Physical limit switches
are positioned at the true safety stops of each end of each stage
so that the supervisory control can prevent the user from making
mistakes that actually damage the hardware.
Fig. 3. Golf Ball Table
 III. SAFETY ENHANCEMENTS
Due to the fact that the user’s remote interaction with the
control hardware allows more freedom of motion, it was decided
that for safety, there exists a need to monitor the status of that
remote connection and disable any effectors if the connection
were to drop. The user connects to the hardware through a VCL-
like server interface where they make a reservation to gain
access to the hardware. The server software is aware of the status
of that reservation, as well as the status of the user’s connection
to a dedicated virtual machine. By using Transmission Control
Protocol (TCP) communication between the server and the
supervisory myRIO, the server is able to make the supervisory
control aware of whatever changes occur to the user’s
connection. When a reservations begins, the server
communicates to the supervisory controller whether the
reservation was for the use of the myRIO or PLC. The
supervisory control then switches to the correct device. Once
established, if the user loses their connection to the virtual
machine, a command is sent to the supervisory control so it can
take action to make the Cartesian robot safe by disabling the
motor outputs. If the user reestablishes connection to the virtual
machine, the supervisory control is again notified so it can re-
enable the motor outputs. Once the user’s reservation ends, the
server communicates this and the supervisory control resets the
experiment in preparation for the next reservation.
IV. PLC UPGRADE
In this design and the previous, a PLC has been used in order
to control the system. However, for a number of reasons, it was
decided to revise the base configuration (and retrofit the elevator
as well) with a different PLC from Allen Bradley, the Micro850,
which is shown in Fig. 4. The drivers of this decision were the
unnecessary complexity of network licensing for RSLogix, and
the lack of native ability to produce pulse train outputs, and the
need to have the supervisory control be able to reset and wipe
the unit. In order to control the 3 degrees of freedom and meet
system requirements, the PLC must be able to create a control
signal for each axis and have a program reset capability.
Another improvement is the software used in order to program
the PLC. Instead of the recurring cost of licensed software, free
versions of commercially available software is used. The unit
selected was from the Micro800 series.
In order to satisfy PLC program reset functionality, an
isolated memory module is used in conjunction with the PLC’s
embedded I/O capability. The module is configured to upload a
blank program upon start-up. Once a user has completed their
session or connection to the server is lost, the Cartesian robot
transitions into a homing sequence. During this sequence, the
Fig. 4. Allen Bradley Micro850
power for the PLC is cycled. This is done in order to ensure that
the next user who is using the PLC will not be able to upload a
completed program done by a user with a prior reservation.
In order to monitor the exact position of the extendable arm,
high speed counters were used in combination with encoders.
The counters used were plug-in modules that provide both
single-ended and differential modes. The isolated high speed
counter module is compatible with two- and three-signal
quadrature, or incremental encoders [6]. The single-ended
incremental option was used to match the encoder configuration.
In order to control the speed at which each axis moves,
individual stepper motors are used. To control these stepper
motors, a pulse train output (PTO) signal with a specific
frequency is utilized. The PLC selected offers individual axis
control by designating 3 PTOs and 3 directional outputs
included in its embedded I/O. This allows for individual control
of the X, Y, and Z axis of the robot. The axes are configured in
the software in order to control the frequency provided to the
stepper motors.
Connected Components Workbench (CCW) is a software
that offers multiple programming styles such as: structured text,
ladder logic, and function block diagrams [7]. For the
application this paper is focused on, ladder logic is used. The
PLC and CCW were designed to be used in parallel so CCW’s
interface compliments the PLC and is easy to use. CCW can be
obtained by downloading it for free online or ordering a free
DVD that includes free support and software updates [7].
Implementing CCW as the programming software used for this
training system decreases the cost for duplicating this training
system because it is offered for free. CCW also has sufficient
documented support including, but not limited to: example code,
function block descriptions, and PLC module configuration
descriptions.
V. OPEN SOURCE
The design and fabrication of the Cartesian robot is intended
to be open source and will be released under a Creative
Commons license so that others will have the ability to replicate
the system if they so desired. Nearly all electrical and
mechanical components used in the construction of the robot
remain commercially available. Exceptions to this are the design
of the augers and the bracketry required when interfacing
components that did not have prefabricated solutions available;
however, 3D models of these items will be made available, as
well as STL files for rapid prototyping.
VI. CONCLUSIONS
The Cartesian robot design presented provides many use
case scenarios related to remote system control of a multiple
DOF system. The training system incorporates two different
control schemes using common programming languages that
will further advance a user’s understanding of remote control
operation. With the upgraded PLC and myRIO, the user will be
able to control the process with either of the two controllers
smoothly. The implementation of this training system also
provides advantageous learning opportunities by offering
students a deeper understanding of how a system with multiple
DOF should operate. With the gripper and golf ball table
assembly, the user will be able to visually obtain direct feedback
for their program. This feedback will allow a teacher and
student the opportunity to practice troubleshooting and problem
solving skills.
VII. FUTURE WORK
In the Spring of 2016, the Cartesian robot and elevator [5]
training systems will be incorporated into the curriculum in
Instrumentation and Practicum classes in the Engineering
Technology program. A variety of projects are planned, starting
with simple placement of balls in a pattern on the table, and
increasing in complexity to vision-based selection of the closest
ball of the desired color; this will allow for the introduction of
other technologies, including path planning. Usage data and user
feedback will be obtained and used to improve the overall
student experience. This data will also be useful in the decisions
about the design of a third generation training system.
ACKNOWLEDGMENT
This workforce product was funded by a grant awarded by
the U.S. Department of Labor’s Employment and Training
Administration. The product was created by the grantee and
does not necessarily reflect the official position of the U.S.
Department of Labor. The U.S. Department of Labor makes no
guarantees, warranties, or assurances of any kind, express or
implied, with respect to such information, including any
information on linked sites and including, but not limited to,
accuracy of the information or its completeness, timeliness,
usefulness, adequacy, continued availability, or owners.
REFERENCES
[1] Williams, W. B., Browne, A. F., “Development of a Remote Laboratory
Architecture for Mission Critical Operations Instruction”, IEEE
SoutheastCon 2016, Norfolk, VA, 30-Mar-3-Apr-2016
[2] Aktan, B., C. A. Bohus, et al. (1996). "Distance learning applied to control
engineering laboratories." IEEE Transactions on Education 39(No.3).
[3] Stefanovic, M., Cvijetkovic, V., Matijevic, M. and Simic, V. (2011), A
LabVIEW-based remote laboratory experiments for control engineering
education. Comput. Appl. Eng. Educ., 19: 538–549.
[4] Bellmunt, O.G.; Miracle, D.M.; Arellano, S.G.; Sumper, A.; Andreu,
A.S., "A distance PLC programming course employing a remote
laboratory based on a flexible manufacturing cell," in Education, IEEE
Transactions on , vol.49, no.2, pp.278-284, May 2006
[5] Aidan Browne, Wesley Williams, Keith Loftus, and Chris Benfield,
“Development of a low-cost electromechanical elevator for
programmable logic and embedded controls training”, 2015 ASEE
Annual Conference and Exposition, Seattle, WA., 15-Jun-2015
[6] http://literature.rockwellautomation.com/idc/groups/literature/document
s/um/2080-um004_-en-e.pdf
[7] http://www.rockwellautomation.com/global/products-
technologies/connected-components/tools/workbench.page