Exploring Autonomous Robot
Carmen Arroyo, Luis Macias
M.E.A.R. is an undergraduate project at the University of Texas at El Paso focused
on developing a six-legged, autonomous walking robot that combines the benefits
of sophisticated motion and adaptive control in a single platform with many
practical applications. The purpose of this project is to build an
autonomous robot that will have the capability of exploring an unknown area and
map key points of interest.
Robots have recently been used to search areas where humans and animals
cannot explore. One application of
robots involves removing the human risk factor from a hostile or uninviting
situation. A potential use for this
design would be in planetary exploration.
Using data taken by robots, scientists can command the unit to go to
rock and soil targets of interest. Initial targets may be close to the landing
sites, but later targets can be further away.
The use of robots in planetary exploration will expand our knowledge of
regions otherwise unapproachable by humans.
Project Description and Operation
A six-legged autonomous walking robot
was built in order to map and explore unknown areas. Mapping the area consists of marking the positions of obstacles
so the robot can distinguish between accessible and inaccessible points. Various sensors are used to determine these
obstacles. The microcontroller accepts
inputs from the sensors in order to recognize land formations and make note of
their existence and location. The
environment was needed in order to simulate the obstacle the robot could
Method of Completion
The objective of the robot is to map an entire area of a predetermined
size. Mapping the area consists of
marking the position of obstacles so that the robot can distinguish between
accessible and inaccessible points.
Various sensors such as sonar and pyroelectric sensors are used to
determine these points. For the
objective, the microcontroller must accept inputs from sensors to recognize
land formations and make note of their existence and location. It must also keep track of the terrain it
has covered in order to precede.
The following block diagram depicts the path of operation that is being
Modular Description of Block Diagram
The environment was needed in order to simulate the kinds of obstacles and
points of interest that the robot could possibly encounter. The environment itself consisted of two
eight foot by four foot by half-inch plywood.
The two boards were placed next to each other in order to create a total
area of sixty-four square feet. Since
the hazardous obstacles needed to be powered by a 120-volt AC power supply
wiring needed to be run. Since we did not want to deal with the problem of
having the robot trip over the wires, a groove was cut so that the wire could
run smoothly under the main area without causing any problems. Then a small
hole was cut in the board allowing the wire to come up and power the light
bulbs that are used.
Points of Interest and
Points of Interest were made out of plastic. The only requirement was that
they be flat. This allows the ultrasonic sensor to detect them more easily.
Hazardous Obstacles consisted of a light bulb with enough wattage to create
enough heat for the pyroelectric sensor to detect. After some testing it was
found that a standard sixty-watt light bulb created enough heat for the sensor
At the beginning of the project the original design called for four
different types of sensors to be used. After the implementation of the ultrasonic
sensor it was found that the other sensors were not needed. But after some
testing and in an effort to correct the inaccuracies in positioning the compass
module was added to the design.
The Devantech Ultrasonic Sensor is the primary sensing unit on the robot.
It uses ultrasonic sound to locate objects in front of it. It has a capability
of detecting objects from three inches to thirty inches in front of it. In our
case we implemented it so that it would wait until an object was five inches in
front. Once it was set off, it then turned on the pyroelectric sensor.
The Eltec Pyroelectric sensor senses differences in the surface temperature
of an object in comparison to the background temperature. It has a window of difference that can be
adjusted using the code in order to make it more or less sensitive.
The Devantech compass module was used in order to better define
position. It uses the magnetic field of
the earth to create a three hundred and sixty degree representation of the
standard compass that we use. We defined zero degrees as north and a clock wise
increment was used to represent each pole. East being approximately ninety
The OOPic provides an Object-oriented language model designed to interact
with the electrical hardware components that you attach to the OOPic. We
designed our hardware interface in software by creating Objects and setting
their properties to define their behavior and interaction with hardware. These
Objects also can be interconnected to form a Virtual Circuit. It can be
programmed using Basic, C, or Java syntax.
Two Scott Edwards Servo Controllers were used to
control the twelve servos needed for leg movement. Controlling servos from a microcontroller is not difficult, but
as the number of servos that need to be controlled increases, the
microcontroller can become bogged down servicing the servos, leaving no time
for monitoring sensors or other I/O. The Serial Servo Controller takes care of
the timing and leaves the microcontroller free to deal with other issues.
HS-422 servos were used for the robot.
They are made up of a DC motor, gear reduction, output shaft with
position feedback, and a control PC board all built into a small rectangular
enclosure. The output shaft is limited to approximately 180 degrees of
rotation, but is normally used in a 90 degree range. The system is
proportional, so the output shaft can be moved quickly or slowly to any
position. These servos are strong,
light weight, reliable, yet easy to setup and control.
Scott Edwards LCD was used in order to display messages which corresponds to
the robots actions. The LCD was also
used to display the positions of the obstacles as they are encountered.
Division of Tasks
The project has many issues dealing with the accuracy of the robot. This means that the robot will not always be
in the position that it is supposed to be in so this leads to many other
errors. These errors include mapping errors and final position errors. In the
mapping errors it is found that since the positioning of the robot is off, the
places where the obstacles and points of interest should have been mapped are
incorrect. Secondly, the final position is also not correct, but in most cases
it is found to be within our specified amount of error, which is one foot. This
means that the robot has to be within one foot of the proposed final position.
Although some problems were encountered throughout the year, the team
worked together and pulled through. At
this point, this prototype is able to identify objects in front of it,
distinguish between points of interest and hazardous obstacles, and complete
the path. This project will be improved
over the years and also used as a motivational tool for upcoming future
This project was sponsored by Benjamin C. Flores Ph.D, Associate Professor of Electrical and
Computer Engineering at the University of Texas at El Paso. At the request
of Dr. Flores this project was pursued in order to develop a physical
attraction that could allure younger students to the fields of engineering and
science. It is also a project that can be enhanced and developed in the future
by other senior project teams.