Transcript Sensors and Vision - Franklin High School

Sensors and Vision
Limit Switches
• Provide information about the state or position
of objects that concern the robot.
• A device activated by contact with an object that
changes the state of its contacts when the object
exerts a certain amount of force.
• The limit switch consists of a body, which
houses and protects the electrical contacts, and
some type of actuator, which moves with
physical contact.
• Actuators can use pressure or rotational motion
to change the contacts of the switch.
• When selecting a limit switch, you will need to
know how much amperage the contacts must
handle, how fast the switch needs to respond,
and how much force the object contacting the
limit switch can generate.
• Each switch takes a certain amount of time to
make (activate) or break (deactivate), so make
sure that the switch can respond fast enough for
the conditions under which you will use it.
• Limit switches are used to prevent the robot
from traveling too far, monitor the doors on
safety cages, confirm raw parts ready for
loading, verify that the machine the robot works
with is in position, and to answer yes/no type
questions.
Proximity Switches
• A solid-state device that uses light, magnetic
fields, or electrostatic fields to detect various
items without the need for physical contact.
• A solid-state device is a unit made up of a solid
piece of material that manipulates the flow of
electrons without any moving parts so it doesn’t
wear out.
• Downsides to solid-state devices:
– They never fully stop the flow of electrons.
– They are vulnerable to electromagnetic pulses and
other strong magnetic fields.
• Main reasons that prox switches fail:
– Physical damage to the switch.
– Voltage that is greater than the prox’s voltage rating.
Inductive Proximity Switch
• Uses an oscillating magnetic field to detect
ferrous metal items.
• When a metal part enters the magnetic field of
the inductive prox, it generates eddy currents in
the part, which takes energy away from the
magnetic field created by the oscillator in the
prox.
Capacitive Proximity Switches
• A switch that generates an electrostatic field;
working on the same principle as capacitors, it
uses the item sensed to complete the capacitive
circuit as well as sense materials at various
distances.
• The capacitive prox will interact with materials
that are both magnetic and nonmagnetic.
• The capacitive prox works just the opposite of
the inductive prox in that the oscillation of the
circuit begins when the item sensed enters the
electrostatic field, and it dies off when no object
is present.
• The field on the inductive prox does not
change, due to the construction of the switch,
while the capacitive prox has an adjustment on
the switch that allows the user to vary the
electrostatic field and change the distance from
which it interacts with objects.
• It is possible to set the switch to detect items
inside cartons, boxes, or other packaging
materials.
Photoelectric Proximity Switch
• Detects levels of light, senses objects via
reflected or blocked beams, and detects colors
via an emitter that sends out a specific
wavelength of light and a receiver that looks for
that wavelength to return after it has interacted
with the surrounding environment.
• With the light level detection photoelectric
switch, we can tune out or ignore the
background, differentiate between parts on a
background, and even detect transparent
objects.
• With proper calibration, this switch could
monitor the empty spaces between parts on a
conveyor line.
• The color detection photoelectric prox switch
analyzes the light coming into the receiver to
determine color or color difference in parts.
• Roboticists have used this sensor to tell the
difference between blue and red balls, follow
colored lines, and carry out other such tasks
involving color recognition.
• Drawbacks:
– Changes in the ambient light will change the light
returning to the switch and the reading, thus
requiring recalibration
– It has a set operating range that is dependent on the
strength of the emitter and the wavelength of the
light sent out.
Tactile and Impact
• Tactile – the ability to sense pressure and
impact
• Tactile sensing is about determining how much
force is being applied, what the shape of the
part is, how it is gripped, if the part is hot or
cold
• Impact – a robot’s contact with an object in the
intended movement path
• Impact sensing is concerned with detection of
collisions, determining if forward movement is
impeded or stopped, and shutting down or
modifying the motion of the system to prevent
damage to whatever is hit and the robot.
• Complex tactile sensors use arrays or organized
groups of sensing elements to gather information
about contact with objects.
• Simple systems consist of elements that have a
digital type of output, either on or off, 1 or 0.
• With this type of sensor, the system knows if it is
gripping the part or not and how many elements
are involved in that grip, denoting the surface over
which the force is applied, but it has no data on the
amount of force the system is using to grip the part.
• High-end complex tactile sensors not only give
information about contact with the part, but also
how much force the system applies to the part.
• Some even measure the temperature of the part
it contacts.
• The common factor in voltage, resistance,
capacitance, and magnetic flux is that change in
the shape of the element causes a change in the
output of that element.
• Calibration is a specified process that ensures
that a precision system performs properly and
provides for any adjustments needed.
• Impact detection play a crucial role in
minimizing or preventing damage to the robot
as well as equipment and people around the
robot.
• Today, we typically monitor the amperage the
motor uses or insert sensing devices designed to
detect impacts.
• Anything that causes the motor to use excessive
amounts of current, such as a too-heavy load,
bad bearings, friction, caked dirt in the joints, or
something else requiring more force, can cause
an impact alarm.
• No matter how we detect the impact, the robot has
to do something to prevent or limit the damage
caused.
• One of the tactics is to E-stop the robot and lock it
in place.
• Robots now disengage the motor from the joints
until the system can come to a full stop to reduce
the force the robot has to stop, reduce the energy
of the impact, and reduce the stress on the robot’s
internal systems.
Position
• Open-loop system – a system that assumes that
everything is working correctly because there is
only limited information on a few positions or no
feedback at all to confirm everything is working as
directed.
• Closed-loop system – a system that sends out the
control pulse to initiate movement and then
receives a return signal that confirms movement,
often including the direction, speed, and distance
moved.
Motor Encoders
• Devices that directly monitor the rotation of a
motor shaft and turn that information into a
meaningful signal.
• The robot controller uses this signal to
determine what the motor is doing as well as
when it has reached the desired point.
• Hall effect sensor - a sensor that uses a
magnetic field to cause voltage flow in a
semiconductor used to track rotation.
Incremental Optical Encoders
• Consist of a disk that has either holes for light
to pass through or special reflectors to return
light, an emitter, a receiver, and some solid-state
devices for signal interpretation and
transmission.
• The transmitter sends out light, and the light
either passes through the holes or reflects back
to the receiver, triggering the electronics of the
encoder to send a signal back to the controller.
• By adding a second row of reflectors or light
windows, offset from the pulse count, and the
appropriate emitter and receiver, we can now
determine the direction of rotation by
comparing the signals from the two rings.
Absolute Optical Encoder
• Adds enough emitters and receivers, usually
four or more, to give each position of the
encoder its own unique binary address.
• When using these encoders, the robot knows
where the motor is at any given point.
• These encoders cost more, have more internal
parts, and require more of the controller’s
processing power.
• The biggest enemies of encoders are oil, metal
chips, shavings, or any other containment that
gets inside the unit and either clouds the disk
that interacts with the light or damages it.
Global Positioning System (GPS)
• A system that determines geographical position
based on the time it takes to receive signals
from three or four separate satellites in orbit
around Earth.
• Mobile robotic vehicles can use GPS to
navigate various terrains or roadways.
• Drones use GPS to make sure that they are on
target and to give their users information about
where they are physically located.
Sound
• For sound detection, a robotic system needs a
microphone, the proper hardware, and the
necessary programming so that the robot can
turn the signal from the microphone into
something useful.
• Patterns of sound can be recorded and used as
triggers for robot actions.
Ultrasonic Sensors
• Robots can use ultrasonic sensors to detect or
“see” objects via the high-frequency sound that
they reflect back.
• An ultrasonic sensor is similar to the photo eyes
or photoelectric prox switches except that they
are emitting and receiving sound instead of
light.
• The emitter sends out a high-frequency sound
pulse, above the human range of hearing, which
strikes objects and then returns to the ultrasonic
sensor.
• The receiver measures the amount of time that it
takes the sound wave to return and uses this to
calculate the distance.
• To determine the distance, take the speed of the
sound wave, multiply it by the time it took for the
wave to return, and then divide by 2.
• Another use for ultrasonic sensors is the
detection of air leaks.
• Small air leaks emit a high-frequency sound that
we cannot hear, but that the ultrasonic detector
can pick up with ease.
Vision Systems
• Allow robots to see the world around them,
using cameras and software that processes the
images taken by the camera.
• Dr. Robert Shillman truly started the odyssey of
robotic vision with the DataMan system in
1982.
• The modern vision system consists of a
specialized light source, a camera mounted on
the robot or at a specific point, and specialized
software from either the robot or the vision
system manufacturer.
• The other key point of a modern vision system
is determining where the part is by the image.
• The basic process is to place a calibration image
in a location where the robot can reach, take a
picture, and then use the system software to
translate that image into point data.
• The next step is to take an image of what the
parts should look like and use this as a
reference image.
• During operation, the robot will take a picture
of the designated area, compare this to the
reference picture, and use the positional data
that it was given at calibration to offset motions.