You are on page 1of 51

ROBOTICS

ME 424 A Unit3

Dr. Ch.Lakshmi Srinivas


Syllabus
UNIT – 3

Robotic sensory devices: Objective, Non-optical position sensors – potentiometers, synchros,


inductocyn, optical position sensors – optic interrupters, optical encoders (absolute &
incremental)

Proximity sensors: Contact type, non contact type – reflected light scanning laser sensors.

Touch & slip sensors: Touch sensors – proximity rod & photo detector sensors, slip sensors –
Forced oscillation slip sensor, interrupted type slip sensors, force and torque sensors

Learning Objectives

After completing this unit, you will be able to:

 Understand the desirable characteristics of sensors


 Identify the types of sensors used in industrial robots
 Classify sensors used in robotic applications
 Know the principle and operation of various sensors such as potentiometric, synchros,
resolvers and inductosyn
 Understand the importance and use of optical interrupters, encoders
 Distinguish between contact type & non-contact type sensors used in robotics
 Understand the significance of slippage effect
 Know the operation of touch, slip and tactile sensors
 Understand the vitality of force and torque sensors in robotic applications
 Know the various considerations in the selection of sensors in industrial robot
applications

ME424U3ChLS201516 1
ROBOTIC SENSORY DEVICES
In robotics, sensors are used for both internal feedback control and external interaction
with the outside environment. Animals and humans have similar distinct sensors. For example,
when you wake up, even before you open your eyes, you know where your extremities are; you
do not have to look to know that your arm is beside you, or that your leg is bent. This is
because neurons in the muscles send signals to the brain, and as they are stretched or relaxed
with the contracting, stretching, or relaxing muscles, the signal changes and the brain
determines the state of each muscle. Similarly, in a robot, as the links and joints move, sensors
such as potentiometers, encoders, and resolvers send signals to the controller, allowing it to
determine joint values. Additionally, as humans and animals possess senses of smell, touch,
taste, hearing, vision, and speech to communicate with the environment. In certain cases, the
sensors may be similar in function to that of humans such as vision, touch, and smell.

Accordingly, the sensors used in robotics are broadly classified as:

Internal state sensors: These sensors deal with the detection of variables such as arm joint
position, velocity or acceleration for robot control. Examples are Potentiometers, Synchros,
resolvers, linear inductive scales, Differential transformers (LVDT, RVDT), Optical
interrupters, Optical encoders (absolute and incremental), Tachometers, and Accelerometers.

External state sensors: Devices used to monitor the robot’s geometric and/or dynamic relation
to its task, environment or the objects that is handling such as range proximity, and touch.
These devices can be of Visual type or non visual type. Examples of non visual types are strain
gauges, pressure transducers, proximity devices, ultrasonic sensors, electromagnetic sensors,
elastomeric materials etc.

There is a huge array of sensors available for measuring almost any phenomenon. These
are typically used for Safety monitoring, interlocking in work cell control, part inspection for
quality control, determining positions and related information about objects in the robot cell.
To choose an appropriate sensor for a particular need, we have to consider a number of
different characteristics. Table 3.1 presents a brief description of these characteristics.

ME424U3ChLS201516 2
Table 3.1 Sensor characteristics

1 Accuracy It is defined as how close the output of the sensor is to the expected value.
If for a given input, the output is expected to be a certain value, accuracy
is related to how close the sensor’s output is to this value.
2 Resolution Resolution is the minimum step size within the range of measurement of
the sensor. In a wire-wound potentiometer, it will be equal to the
resistance of one turn of the wire. In a digital device with n bits, the
resolution will be : =
3 Sensitivity Sensitivity is the ratio of a change in output in response to a change in
input. Highly sensitive sensors will show larger fluctuations in output as
a result of fluctuations in input, including noise.
4 Interfacing Sensors must be interfaced with other devices such as microprocessors
and controllers. The interfacing between the sensor and the device can
become an important issue if they do not match or if other add-on
components and circuits become necessary.
5 Type of The output of a sensor may be digital or analog and, depending on the
output application, this output may be used directly or have to be converted.

(digital or
analog)
6 Cost The cost of a sensor is an important consideration, especially when many
sensors are needed for one machine. However, the cost must be balanced
with other requirements of the design such as reliability, importance of
the data they provide, accuracy, life and so on.
7 Size Depending on the application of the sensor, the size may be of primary
importance.
8 Weight Since robots are dynamic machines, the weight of a sensor is very
important. A heavy sensor adds to the inertia of the arm and reduces its
overall payload.
9 Response Response time is the time that a sensor’s output requires to reach a certain
time percentage of the total change. It is usually expressed in percentage of
total change, such as 95%.
10 Reliability Reliability is the ratio of how many times a system operates properly,
divided by how many times it is used. For continuous, satisfactory
operation it is necessary to choose reliable sensors that last a long time
while considering the cost and other requirements.

ME424U3ChLS201516 3
NON-OPTICAL POSITION SENSORS

Position sensors are used to measure displacements, both rotary and linear as well as
movements. In many cases, such as in encoders, the position information may also be used to
calculate velocities.
Potentiometers
Potentiometers are analog devices whose output voltage is proportional to the position
of a wiper. Figure 3.1 illustrates a typical pot. A voltage is applied across the resistive element.
The voltage between the wiper and ground is proportional to the ratio of the resistance on one
side of the wiper to the total resistance of the resistive element.

Essentially the pot acts as voltage divider network. That is, the voltage across the
resistive element is divided into two parts by a wiper. Measuring this voltage gives the position
of the wiper.

Figure 3.1 Principle of operation of (a) Rotary POT and (b) linear POT

As shown in figure 3.1, the wiper makes physical contact with wires on the resistive
coil. Point ‘a’ in the figure corresponds to zero output (i.e. zero resistance). In the rotary
design, the output resistance is proportional to θ, while in linear case it is‘d’.

ME424U3ChLS201516 4
Here the pot acts as a voltage divider, the output will be proportional to the resistance as
Vout = Vs

For the pot, it is important that the resistance r be linearly related to the angular distance
traveled by the wiper shaft. Although, it is possible to obtain pots that are nominally linear,
there is always some deviation from linearity as shown in figure 3.2.

Figure 3.2 Output of a pot showing the theoretical linear and actual non linear
characteristics as a function of angular position ‘θ ‘

Generally, the nonlinearity of a pot is defined as the maximum deviation ε from the ideal
straight line compared to the full scale output. Vmax

ε
% Non- Linearity = × 100

The inevitable presence of this nonlinearity is the main drawback of pot.

ME424U3ChLS201516 5
Figure 3.3 Thin film resistive POTS

The thin film resistive POT shown in figure 3.3, has much better resolution than the
wire wound version. Also, contact noise problems are reduced.
Resolution (smallest change in wiper position that can be sensed) of a wire wound pot is
given by Vs/N. [A change in Vout will occur, only when the sliding contact moves from one coil
loop to the adjacent one]

% Resolution = 100 = 100/N

Potentiometers are either wire wound or thin film deposit (also called conductive
plastic), which is a deposit of a thin film of resistive material on a surface. The major benefit of
thin film potentiometers is that their output is continuous and thus less noisy. As a result it is
possible to electronically differentiate the output of this type of resistor to find velocity.
The major draw backs of pots are noise, wear and tear, heating of coil, prone to dust and dirt.
Example Problem 3.1
A wire wound potentiometer is to be used to measure angular position. A 100 turn (i.e. 100
loops) resistive element is used and the wiper can rotate 300o. 10 V dc is applied to the pot.
Determine the resolution of the device.

Solution: ∆ = = 0.1 /

The % resolution is therefore,

.
% = 100 = 1.0

ME424U3ChLS201516 6
SYNCHRO
A significant practical problem with the POT is that it requires a physical contact in order to
produce an output. This difficulty is avoided in this transducer. Synchro is a rotary transducer
that converts angular displacement in to a voltage or an ac voltage in to angular displacement.

Machine tool and robotic manufacturers have increasingly turned to resolvers and synchros
to provide accurate angular and rotational information. These devices excel in demanding
factory applications requiring small size, long term reliability, absolute position measurement,
high accuracy and low noise operation.

NOTE: Historically, this device was used extensively during World War II, but
technological innovations that produced other position sensing elements caused it to fall
from favor. In recent years, however, advances in solid-state technology have again made
the synchro a possible alternative for certain types of systems, among them are robots.

Normally, a synchro system is made up of a number of separate three phase components.


[e.g., the control transmitter (CX), control transformer (CT), and control differential
transformer (CDX)]. These elements all work on essentially the principle of the rotating
transformer.

Typically there are two types of synchro systems, one is two element systems and the
other is three element systems. These are used to measure angular position or the difference
between this and a command position (i.e. position error)

ME424U3ChLS201516 7
TWO ELEMENT SYNCHRO SYTEM

Figure 3.4 a two- element (control transmitter CX &control transformer CT) synchro
system used to measure angular displacement.
(θ is the relative angle between CX and CT)

Here an ac voltage is applied to the rotor of the CX and that the wye- configured stators of
the CT and CX are connected in parallel.

Using elementary transformer theory, it can be shown that the magnitude of the transformer
rotor voltage Vout (t) is dependent on the relative angle θ between the rotors of the CX and CT.
In particular, this output voltage is

Vout (t) =Vm Sin θ Sin ωact --------- (1)

Where, Vm and ωac are the amplitude and radian frequency of the reference (or “carrier”) ac
voltage.

It can be seen that the output voltage has its maximum magnitude when the two rotors
are at right angles to one another and that it is zero when they are at either parallel or anti-
parallel. As a consequence, the CT is sometimes referred to as “null detector”.

ME424U3ChLS201516 8
THREE ELEMENT SYNCHRO SYSTEM

Figure 3.5 an example of a servo using three- element synchro system

Figure 3.5 shows a typical servo control system using a three element synchro used in sheet
rolling machines. To maintain a uniform product (e.g., uniform steel sheet thickness), the slave
roller’s speed ωS (t) must be synchronized to that of the master, i.e., ωM (t). The CDX is used to
provide the desired angular relationship between the master and slave. The output signal of the
CT is the difference between this desired and the actual master-to-slave angles, i.e., the error,
and is used to provide the slave motor drive signal. The angular relationship between the master
and slave rollers can be adjusted during the running of the process by rotating the shaft of the
CDX.
Ideally, the ac signals from the CX are in phase with those produced at the CT. However,
physical differences in the structures of the two devices that are inevitably present produce
phase shifts that may be undesirable. A synchro control differential transformer (CDX) is
sometimes used to adjust the phase shift between the two synchro units.

ME424U3ChLS201516 9
APPLICATION OF TWO ELEMNT SYNCHRO SYSTEM

Figure 3.6 a synchro used in a position servo loop

A typical position servo control application is shown in figure 3.6. The desired angular
position is θ1 where as θ2 is the actual angular position of the motor shaft. VR is the ac reference
(carrier) voltage. It is observed that the command or input (θ1) will produce a command voltage
from the CX. The CT will then produce an error voltage in accordance with eqn (1).

Where angular error θ = θ1- θ2. This error signal is amplified and causes the servomotor to
rotate until θ is again zero.

In such type of applications the two element synchro provides a rugged, reliable, and cost
effective method of monitoring position error.

From the above we observe that because of the need to convert the command position in to
a physical angular rotation of the CX rotor, such a system is not always practical in applications
requiring the interfacing to digital devices. To overcome this problem, a hybrid system shown
in figure 3.7 is used. With this system interfacing with digital system is quite convenient.

ME424U3ChLS201516 10
Figure 3.7 a typical servo control system using two element synchro in which CX is replaced by
a digital-to-servo converter

The digital-to-servo (D/S) converter shown in figure 3.7 replaces the CX of the position
servo system shown in figure 3.6. A digital position command signal from a computer (e.g.
master) is transformed in to a three-phase ac voltage by the D/S converter. (This voltage
corresponds to that produced by the CX due to a physical rotation of θ1).The CT is then acts a
position error sensor and the system behaves in a manner that is identical to that figure 3.6.

ME424U3ChLS201516 11
RESOLVERS

A Resolver is actually a form of synchro and for this reason it is often called a “synchro
resolver”. One of the main differences between the two devices is that the stator and rotor
windings of the resolver are displaced mechanically 900 to each other instead of 1200 as is the
case with synchro.

Figure 3.8 Schematic diagram of a Simple resolver electrical circuit

Figure 3.8 shows the most common form of electrical circuit of a resolver, which has a
single rotor and two stator windings.

With the rotor excited by an ac carrier voltage V Sin ωact, the two stator voltages become

V1-3(t) = Vsinθ Sin ωact


V2-4(t) = Vcosθ Sin ωact
Where, θ is the resolver shaft angle.

It is used in much same way as the synchro CX to monitor shaft angle.

ME424U3ChLS201516 12
Figure 3.9 Resolver transmitter connected to a resolver control transformer

Another form of a resolver has two stator and two rotor windings as shown in figure 3.9. In
actual use, the carrier voltage is applied to any one and the other one is normally shorted.

To utilize a resolver in a servo system, it is necessary to employ two resolvers in much the
same way as was done with the synchro system.

Figure 3.9 shows a resolver transmitter (RX) and resolver control transformer (RT) in a
simple position servo. Here RX and RT are used to obtain the difference between the actual and
desired angles (i.e.θ1-θ2). It is important to understand that, although angular position can be
monitored using a single resolver, this is usually not done in servo-controlled devices, because
of the need to utilize an error signal to drive the system actuator.

ME424U3ChLS201516 13
APPLICATION OF A RESOVER
(SERVO POSITION CONTROL SYSTEM)

A position servo system utilizing D/R converter (or DRC) is shown in figure 3.10. Here
the output of the DRC is an ac voltage; it uses an ac amplifier, together with a phase-sensitive
detector and integrator to obtain the appropriate drive signal to the servo amplifier. In this case
the command signal is digital, where as the monitored position (and the error) is analog signal.

Figure 3.10 a position servo that uses a D/R converter

ME424U3ChLS201516 14
DIFFERENCE BETWEEN SYNCHRO & RESOVER

Figure 3.11 Stator /Rotor set up for synchro and Resolver

SYNCHRO RESOLVER
They employ single-winding rotor that They employ single-winding rotor that
revolve inside fixed stator. revolve inside fixed stator.

The stator has three windings oriented 120º The stator has two windings oriented at
apart and electrically connected in a Y- 90º apart.
connection.
If the rotor of a synchro is excited with a If the rotor of a resolver is excited with a
reference voltage, Vsinωt, across its reference voltage, Vsinωt, across its
terminals R1 and R2, then the stator's terminals R1 and R2, then the stator's
terminal will see voltages in the form: terminal will see voltages in the form
1 3= sin sin 1 3= sin sin )
3 2= sin sin( + 120 ) 4 2= sin sin( + 90 )
2 1= sin sin( + 240 ) = sin sin cos )
Where θ is the shaft angle. Where θ is the shaft angle.
As synchros have three stator coils in a These are less costly as the stator coils are
1200 orientation, they are more difficult to oriented in 900, they are easier to
manufacture and hence costly. manufacture.

ME424U3ChLS201516 15
INDUCTOSYNS
Synchros and resolvers inherently measure rotary position, but they can make linear position
measurements when used with lead screws. An alternative, the Inductosyn™ (registered
trademark of Farrand Controls, Inc.) measures linear position directly. In addition, Inductosyns
are accurate and rugged, well-suited to severe industrial environments, and do not require
ohmic contact.

Figure 3.12 Linear Inductosyn

The linear Inductosyn consists of two magnetically coupled parts; it resembles a multipole
resolver in its operation (see Figure 3.12). One part, the scale, is fixed (e.g. with epoxy) to one
axis, such as a machine tool bed. The other part, the slider, moves along the scale in
conjunction with the device to be positioned (for example, the machine tool carrier).

The scale is constructed of a base material such as steel, stainless steel, aluminum, or a tape of
spring steel, covered by an insulating layer. Bonded to this is a printed-circuit trace, in the form
of a continuous rectangular waveform pattern. The pattern typically has a cyclic pitch of 0.1
inch, 0.2 inch, or 2 millimeters. The slider, about 4 inches long, has two separate but identical
printed circuit traces bonded to the surface that faces the scale. These two traces have a
waveform pattern with exactly the same cyclic pitch as the waveform on the scale, but one

ME424U3ChLS201516 16
trace is shifted one-quarter of a cycle relative to the other. The slider and the scale remain
separated by a small air gap of about 0.007 inch.

Inductosyn operation resembles that of a resolver. When the scale is energized with a sine
wave, this voltage couples to the two slider windings, inducing voltages proportional to the sine
and cosine of the slider's spacing within the cyclic pitch of the scale. If S is the distance
between pitches, and X is the slider displacement within a pitch, and the scale is energized with
a voltage V sinωt, then the slider windings will see terminal voltages of:

2
( )= sin sin

2
( )= sin cos

As the slider moves the distance of the scale pitch, the voltages produced by the two slider
windings are similar to those produced by a resolver rotating through 360º. The absolute
orientation of the Inductosyn is determined by counting successive pitches in either direction
from an established starting point.

Rotary inductosyns can be created by printing the scale on a circular rotor and the slider's track
pattern on a circular stator. Such rotary devices can achieve very high resolutions.

OPTICAL POSITION SENSORS

Some sensors utilize optical hardware and techniques are frequently used to perform the
position determination task with relative ease and good accuracy. We now discuss such sensors
and their application to robotics.

1. Opto-Interrupters
2. Optical Encoders
a. Rotary absolute encoders
b. Optical incremental encoders

ME424U3ChLS201516 17
OPTO-INTERRUPTER

In PTP type robots the beginning and end points should be accurately determined, the
actual path between these points is not important. Hence little or no position information is
utilized by the robot’s control system except the trajectory end point.

The actuators drive the joints of the robot until the final position is sensed, at which time
the actuating signals are removed. In effect, an open loop control scheme is used. Programming
is accomplished by moving the endpoint sensors to different locations.

In such applications a mechanical switch or micro switch is an ideal device. The main
drawbacks of these switches are interfacing the switch with the microprocessor, inevitable
contact bounce problem and limited life expectancy. (Due to mechanical wear and tear)

An optical technique can be used to produce the required ability to sense “end of travel” by
the device called an opto-interrupter.

Figure 3.13 Simple opto-interrupter showing light emitter-receiver assembly and disk with flag

ME424U3ChLS201516 18
As shown in figure 3.13, a transparent disk with at least one dark sector is placed
between a light emitter (e.g., an LED) and a light receiver or sensor (e.g., a phototransistor).

Light will reach the receiver until rotation of the disk causes the “black flag” to block it.
A binary or “on-off” signal can be generated and used to sense the endpoint of travel. For
example, the output (i.e., the collector) of the phototransistor will be low as long as light
impinges on the transistor’s base. On the other hand the collector voltage will be high when
there is no light.

Figure 3.14 Block diagram of a simple unidirectional motor control circuit. the motor
begins to rotate when the switch is closed

The block diagram of a simple electronic circuit that makes use of opto-interrupter sensor to
drive a robot axis to the end of travel is shown in figure 3.14. Here the system is actuated by
momentarily closing the start switch. The motor will continue to rotate until the black flag on
the disk prevents light from reaching the light sensor. When this occurs, the motor voltage is
turned off and the axis coasts to a stop.

ME424U3ChLS201516 19
Figure 3.15 Possible realization of a sensor and logic circuits for simple motor control

A possible realization of the logic and sensor electronics is shown in figure 3.15. Initially,
the black flag on the disk is initially placed in the slot of the opto-interrupter; the collector of
the phototransistor will be about 5V, so that S1 will be high. If the one-shot and debounce
circuit is designed so that its output is normally high and goes low only when the one-shot is
triggered by the start switch is being grounded, S2 will normally be high also. Therefore, the
signal to the motor drive circuitry is low and the motor does not turn. It is important to note that
this simple circuit permits only unidirectional rotation of the motor.

The wave forms of the digital signals S1, S2, and S3 are shown in figure 3.16.

Figure 3.16 Timing diagram for the digital signals associated with the sensor
and logic circuit of motor control system

ME424U3ChLS201516 20
OPTICAL ENCODERS
One of the most widely used position sensors is the optical encoder. Capable of
resolutions that are more than adequate for robotic applications, these noncontact sensory
devices come in two distinct classes.

(a) Absolute
(b) Incremental

In the case of absolute, the encoder is able to give the actual linear or rotational position
if power has just been applied to the electromechanical system using the sensor. Thus a robot
joint equipped with an absolute encoder will not require any calibration cycle since the
controller will immediately, upon power-up, know the actual joint position.

In the case of incremental encoder, the encoder provides positional information relative
to some reference point. A robot utilizing an incremental encoder must, therefore, first execute
a calibration sequence before true positional information can be obtained.

Although either linear or rotary encoders for both of the foregoing classes are available,
the rotary device is almost exclusively used in robotic applications. One of the most important
reasons for this is that revolute joints far outnumber prismatic ones in robots currently being
manufactured. As such linear encoders are much costlier than rotary encoders

ME424U3ChLS201516 21
ROTARY ABSOLUTE ENCODERS

The absolute encoder is capable of giving the correct rotary position at all times even after
power-up has occurred. The device produces a separate and unique coded word for each shaft
position.

A major advantage of the absolute encoder is that even if system power is accidentally lost
(due to a power outage or relay trip, for example) the device will remember where it is and will
report this to the system as soon as power is restored.

Absolute encoders usually consists of three major elements:


1. A multiple-track (or channel) light source
2. A multiple-channel light receiver
3. A multiple-track rotary disk

Figure 3.17 schematic arrangement of absolute encoder including the disk, light sources, light
detectors, and electronics.

ME424U3ChLS201516 22
Figure 3.18 a four-track, sixteen-sector pure binary coded disk

Normally, light emanating from a linear, N-element light source (e.g., LEDs) is made to
pass at right angles through the disk and is received (or collected) by a corresponding linear
array of N light sensors (e.g., phototransistors) mounted on the opposite side of the disk. (See
Figure 3.18)
The disk is divided in to circumferential tracks and radial sectors. Absolute rotational
information is obtained by utilizing one of several possible code formats. For example, Figure
3.18 shows a four track 16-sector pure binary coded disk. [The other coding schemes are BCD
and gray code]

It can be seen from Figure 3.18 that the resolution of the disk is 22.50 (360/16) since one
complete disk revolution is 3600 and there are 16 sectors. [Assume shaded areas represent a
binary 1 and clear areas a binary 0].

ME424U3ChLS201516 23
For example, if sector 11 is in the region of the LEDs the output of the photo transistors
will be 1011 or decimal 11. It is clear from this discussion that the absolute disk position is
known simply by reading the photo-detector outputs.

In practice, it is possible to produce absolute encoders with up to 13 separate channels (i.e.,


13 bits) which means that resolutions up to 360/213 = 0.0440 are possible for a single complete
rotation of the disk.

Table 3.2 Binary and Gray codes

Table 3.2 shows the difference between a binary code and a gray code. In the binary
code system, there are many instances where more than one set of bits change sign
simultaneously, where in gray code, at any particular location, there is always only one bit-
change to go back or forth. The importance of this difference is that in digital measurements,
unlike popular perception, the values of signals are not constantly read, but the signal is
measured (sampled) and held until the next sample reading. In binary code, where multiple bits
change simultaneously, if all changes do not happen exactly at the same time, they may not all
register. In gray code, since there is only one change, the system will always find it.

ME424U3ChLS201516 24
Figure 3.19 a four-bit Gray coded disk

INCREMENTAL ENCODER
Optical incremental encoders are widely used to monitor joint position on robots. In
addition, they are the sensor of choice in a variety of machine tools, including lathes, x-y tables,
and electronic chip wire and hybrid die bonders. The major reason is that they are capable of
producing excellent resolution at a significantly lower cost. It is important to understand that if
power is accidentally lost during an operation, calibration must be performed again since the
incremental encoder has no memory.

In this type of encoder, the areas (arcs) of opaque and transparent sections are equal and
repeating. Since all arcs are the same size, each represents an equal angle of rotation. If the
number of divisions (arcs) increases, the accuracy increases as well.

Typical incremental encoders can have 100, 128, 200, 256, 500, 512, 1000, 1024, 2000, and
2048 arcs reporting angular displacements with a resolution of 3.60 to 0.17580.

Optical encoders are either opaque disk with the material removed for transparent areas or
are clear material like glass with printed opaque area. Many encoder disks are also etched, such
that they either reflect the light or do not reflect the light. In that case the light source and the
pick-up sensor are both on the same side of the disk.

ME424U3ChLS201516 25
Figure 3.20 (a) a simple rotary incremental encoder wheel
(b) a linear incremental encoder

An incremental encoder acts as an integrator, because the controller actually counts the
number of signals the encoder sends, determining the total positional change, and consequently
it is integrating the position signal.

Many photo-detectors are analog devices. This means that as the magnitude of the light
varies, their output varies too. Therefore, as one section on the encoder disk approaches the
detector and the projected light intensity increases to a maximum, the output of the detector
rises before falling again as departs. Consequently, a squaring circuit is used to condition the
signal. Figure 3.21 shows the TTL output of an incremental encoder.

Figure 3.21 TTL outputs of the A and B channels of an incremental encoder


(a)A leads B, when clockwise rotation occurs (b) A lags B, when counter-clockwise
rotation occurs

ME424U3ChLS201516 26
If only one set of slots is used, it will be impossible to determine whether the disk is
rotating clockwise or counterclockwise. In order to determine the type of rotation, encoder
disks have two sets of slots (two channels), ½ step out of phase with each other (figure 3.21).
As a result, the output signals of the two sets of slots are also a ½ step out of phase with each
other. The controller can compare the two signals and determine which one changes from high
to low or vice versa before the other signal. Through this comparison, it is possible to
determine the direction of rotation of the disk.

As shown in figure 3.22, by counting both the leading edges as well as the trailing edges
of the output signals of the encoders on both channels, it is actually possible to increase
resolution of the output of incremental encoders without increasing the number of slots.

Figure 3.22 output signals from incremental encoder

ME424U3ChLS201516 27
PROXIMITY SENSORS

A proximity sensor is used to determine that an object is close to another object before
contact is made. There are many different types of proximity sensors such as magnetic, eddy
current, Hall Effect, optical, ultrasonic, inductive and capacitive. These sensors can be either
contact type or non contact type.

CONTACT PROXIMITY SENSORS:

Figure 3.23 Simple contact rod proximity sensor mounted on one finger of a robotic
gripper

Figure 3.23 shows a simple contact rod proximity sensor mounted on one finger of a robotic
gripper. It consists of a rod that protrudes from one end and a switch or other linear position
monitoring element located within the body of the sensor. As the robotic manipulator moves,
the sensor will become active only when the rod comes in contact with an object or an
obstruction. When this occurs, the switch mounted inside the sensor will close (or open, if that
is more convenient). The change of state of the switch, monitored through the robot’s I/O
interface, will cause an appropriate action to take place.

If the simple on-off switch is replaced by any of the linear position sensing devices such as
pot, LVDT etc, the sensor can detect the actual position of the object or obstacle.

ME424U3ChLS201516 28
For example if the contact rod of the above sensor is attached to the magnetic core of an
LVDT or wiper of pot, the motion will be converted in to a voltage that will be proportional to
the actual distance of the end of the rod (and hence the object) from some reference point on
the robot (e.g., the end of the gripper). In addition, the approach velocity can be obtained from
this information by performing either an analog or digital differentiation. Thus both distance
and approach velocity can be monitored using such a contact sensor.

NON-CONTACT PROXIMITY SENSORS:


These sensors do not require any physical contact with the object or obstacle in order to
produce a signal that can be used by a robot. We discuss here only two such sensors: (i)
Reflected light sensors and (ii) Scanning laser sensors.

REFLECTED LIGHT SENSORS


One of the simplest types of proximity sensors that uses light reflected from an object
and has been used experimentally on a robot gripper is shown in figure 3.24.

Figure 3.24 Reflected light sensor: light source-detector assembly

The sensor consists of a source of light and a photo-detector separated by about 8 mm and
tilted symmetrically toward one another. This, together with lenses mounted in front of the
assembly, produces focused incident and reflected beams.

Figure 3.25 (a) shows the photo-detector voltage as a function of object distance from the
(detector) lens. Figure 3.25 (b) shows the placement of such sensors on a robotic gripper. In
this way, proximity in several directions can be monitored simultaneously. (e.g., ahead and
below the robot’s hand)

ME424U3ChLS201516 29
Figure 3.25 Reflected light sensor: (a) detector output voltage as a function of distance
(b) two-dimensional sensor array mounted on a parallel jaws gripper

The maximum sensor output will occur, when an object is at the focal point. However, there
are several difficulties with this type of sensor.

Difficulty 1: As shown in 3.25 (a), two different object positions produce the same voltage,
except when the object is at the focal point. Since a one-to-one correspondence between
position and detector voltage does not exist, additional logic or hardware is required to
eliminate the ambiguity. For example, if the robot is moving and the sensor signal is increasing,
it is clear that the object is on the far side of the focal point (i.e., has yet to reach this point, and
so the output corresponds to the larger of the two position values). If, however, the signal is
decreasing, the focal point has been passed and the smaller distance should be used. Another
way to avoid this ambiguity is to place several sensors at different angles.

Difficulty 2: ambient light will shift the curve in figure 3.25 (a) up or down depending on the
intensity. It can be avoided by pulsing the light frequency at about 6 kHz rate.

Difficulty 3: a more difficult and perhaps impossible problem to overcome is the sensor is
sensitive to the reflectivity of the object or obstacle.

It would be difficult to use this sensor for absolute position measurement because of its
sensitivity to variations in light source output, drift in the detector characteristics (due to
ambient temperature fluctuation) and environmentally caused changes in reflectivity of the
object.

ME424U3ChLS201516 30
SCANNING LASER SENSOR

This is a commercially available and costlier sensor. It consists of laser light source, two
mirrors, one of which is rotated by an ac motor, and a lens-photo-receiver assembly as shown
in figure 3.26.

Figure 3.26 schematic diagram of scanning laser proximity sensor

This scanning laser device has been used to permit an industrial robot to arc-weld curved
objects. The incident light beam from the laser (helium-neon) is swept across the object surface
by the action of the motor driven triangular mirror. Note that this occurs three times for each
motor revolution. A lens mounted in front of a photo-detector (e.g., a phototransistor) permits
light reflected from only one point on the object’s surface to be acquired. Distance from the
sensor to this point is determined by synchronizing the ac motor voltage with a high frequency
clock. The number of clock pulses from the time this voltage is zero until the photo detector
receives reflected light is a measure of the distance.

Tracking (in the case of the welding application, for example) is achieved by mounting the
sensor on the end effector of the robot, thereby allowing the entire sensor to be moved to
different locations in space. Black, transparent, or extremely shiny objects cause problems for
this proximity technique.

ME424U3ChLS201516 31
TOUCH AND SLIP SENSORS

TOUCH SENSORS

Touch sensors are devices that send a signal when physical contact has been made. The
simplest form of a touch sensor is a micro switch, which either turns on or off as contact is
made. The micro switch can be set up for different sensitivities and ranges of motion.

For example a strategically placed micro switch can send a signal to the controller if a
mobile robot reaches an obstacle during navigation. More sophisticated touch sensors may send
additional information, for example a force sensor provides not only touch information but also
provide magnitude of force.

Figure 3.27 operation of a simple touch sensor

ME424U3ChLS201516 32
TACTILE SENSORS

The term tactile sensing does not have a universally accepted meaning. Harmon has
defined it to be “the continuous-variable sensing of forces in an array.

An ideal tactile sensor should have the following attributes:

1. Tactile sensors should be compliant and rugged. (i.e., durable in the manufacturing
environment)
2. Sensors (i.e., sensory arrays) should be ‘smart’ meaning that they should process
most of the information before communicating with the robot.
3. Sensor resolution should be on the order of 100 mils, although some applications
could require a larger or smaller value.
4. The sensors should be able to detect forces as low as 5 to 10 gm, with a dynamic
range of about 1000 to 1.
5. Sensor response should be stable (with time), monotonic (and preferably linear,
although some nonlinearity would be acceptable), and most important, not exhibit
hysteresis.

A Tactile sensor is a collection of touch sensors which, in addition to determining


contact, can also provide additional information about the object. This additional information
may be about the shape, size, or type of material. In most cases, a number of touch sensors are
arranged in an array or matrix form.

ME424U3ChLS201516 33
Proximity rod tactile sensor
It consists of an array (i.e., 4 x 4 = 16) of single contact rod proximity sensors. The
robot wrist on which the device mounted be moved down toward and parallel to the table or
other surface on which an object is resting. The main difficulty with this technique is that how
much the robot descends toward table surface.

Figure 3.28 Three-dimensional proximity rod tactile sensor


(a) Top view (b) Side view (c) Sensor descending (d) Sensor in partial contact
with object (e) Sensor in full contact with object

As the tactile sensor comes in contact with an object, depending on the shape and size
of the object, different touch sensors react differently at a different order. This information is
then used by the controller to determine the size and shape of the object.

Figure 3.29 a tactile sensor can provide information about the object
Figure 3.29 shows three simple set-ups, one touching a cube, one touching a cylinder,
and one touching an arbitrary object. As can be seen, each object creates a different unique
signature that can be used for detection.

ME424U3ChLS201516 34
Photo-detector tactile sensor
Figure 3.30 shows the construction details of a typical photo-detector tactile sensor. The
portion of the sensor that comes in contact with the object to be sensed is covered with an
elastomer (a rubber like material). In addition a piece of this material extends through the
sensor structure. Mounted on the back of the body of the device is a photo emitter-detector
assembly. When object comes in contact with the touch surface, if the elastomer is compressed
a minimum distance, the material extending through the body breaks the beam of the photo
sensor. A thresholding circuit is used to provide the binary information about the object. From
this sensor, objects two dimensional information such as size, shape, and part orientation can be
obtained. It consists of an array of (8 x 8 for Lord LTS 100 model) such sensors. The major
difficulties of the photodetector tactile sensor are hysteresis effects of elastomer material and
wear and tear of the elastomer material.

Figure 3.30 construction details of a typical photo detector tactile sensor

Figure 3.31 working details of a photo-detector tactile sensor

ME424U3ChLS201516 35
Skin like tactile Sensor

Attempts have also been made to create somewhat of a continuous skin-like tactile
sensor that could function similarly to human skin. In most cases, the design revolves around a
matrix of sensors embedded between two polymer type layers, separated by a mesh as shown in
figure 3.32.

Figure 3.32 Skin-like tactile sensor

As a force is applied to the polymer, it is distributed between a few surrounding sensors,


where each one sends a signal proportional to the force applied to it. A microprocessor reads
the sensors sequentially in order to determine the shape of the object and the contact force at
each location.

ME424U3ChLS201516 36
SLIP SENSORS
One of the capabilities of the human hand is its ability to determine when an object that
is being grasped is slipping. The biological control system associated with the hand utilizes the
inputs from the appropriate slip receptors and causes the gripping force to be increased or
decreased, as the case may be.
Machine determination of slippage of a part or object when in the grasp of a robot or other
electromechanical hand is still under active research. The need is to detect slip rapidly and to
adjust the gripping force “on the fly” to prevent the part from being damaged in a fall.

Lift and try technique


Lift and try technique uses motor current of a particular joint or set of joints on a robot
as a measure of whether or not a part is slipping. Current monitoring can be either digitally or
analog manner. Here the gripper is first oriented correctly, next placed over the particular part,
and then a certain minimum grasping force applied. If no increase in motor current is detected,
the manipulator is commanded to return to the starting point. The force is then incremented by
some predetermined amount and the robot “tries again”. The procedure is repeated until the
monitored joint current does increase, at which time it is assumed that the part is not slipping
and is properly grasped.

Figure 3.33 lift and try technique for slip detection

ME424U3ChLS201516 37
Drawbacks of this technique of slip sensing are even if the part is successfully raised above
the resting surface, there is no guarantee that it will not slip out of the gripper as the
manipulator moves. Another problem is that if a fragile part is to be lifted, the minimum
applied gripping force should be small, to avoid crushing.
In addition to the lift and try procedure, a number of experimental devices based on optical,
magnetic, or conductive sensing techniques have been developed.

Forced oscillation slip sensors

This is one of methods developed for detection of slippage. Forced oscillation was used
whereby any translation of the part in a direction tangential to the surface of the gripper jaws
(i.e., orthogonal to the direction of the applied gripping force) caused a short burst of voltage
(i.e., a “spike”) to be generated.

Figure 3.34 prototype slip detector using a sapphire needle and piezoelectric crystal to
produce a forced oscillation.

As shown in figure 3.34, a sapphire needle protrudes from the surface of the sensor and is
in contact with the object being grasped. If the part begins to slip, the needle will be displaced
and will produce mechanical deformation in a piezoelectric crystal (e.g., Rochelle salt). The
resultant generated voltage spike can be sensed using a threshold detector and the gripping
force increased incrementally until the part stops slipping.

ME424U3ChLS201516 38
The main drawback of the above device is the rubber damper is sensitive to non-slip related
motion. It is important to note that such a detector must be able to sense slippage accurately
while a part is being moved. That is it is necessary to reduce as much possible this sensitivity to
non-slip related motions. Another problem with the above slip sensor is wear of the needle and
periodic replacement of the same.

The main difficulties of the above are eliminated in the sensor shown in figure 3.35.
The sapphire needle was replaced by a more robust steel ball (0.5 mm in diameter). The
crystal transducer is replaced by permanent magnet-coil arrangement. Any motion of a part
along the gripper’s jaws will cause a mechanical displacement of the coil which will in turn
produce a voltage output. Both rubber damper and oil damper and thresholding circuitry will
reduce the sensitivity of the sensor to manipulator vibrations (other than slippage).

Figure 3.35 improved forced oscillation slip sensor

ME424U3ChLS201516 39
Interrupter type slip sensors
The first one, magnetic roller type sip sensor uses magnetic head such as one in a tape
recorder as a transducer (figure 3.36). The roller has a permanent magnet embedded in it in one
spot. In its reset position, the magnet is directly over the magnetic head, but when a slip occurs,
the roller rotates the magnet away from the head. Unfortunately, if a second slip occurs before
the sensor is reset it will go undetected.

Figure 3.36 Roller type slip sensor with magnetic transducer


The second one, roller type slip sensor (figure 3.37) has an optical transducer. A slit
made in the roller allows light from a lamp or an LED to pass through the roller to a
photodiode. A slipping item rotates the slit away from the lamp and photo detector blocking the
detector from the light source. Multiple slips also trouble this sensor, but the problem can be
solved by cutting multiple slits in the roller.

Figure 3.37 Roller type slip sensor with optical transducer

ME424U3ChLS201516 40
FORCE AND TORQUE SENSORS
FORCE SENSORS
The capacity to measure forces permits the robot to perform a number of tasks. These
include the capacity to grasp parts of different sizes in material handling, machine loading, and
assembly work, applying the appropriate level of force for the given part. Force sensing in
robotics can be accomplished in several ways.

The various methods are:


1. Piezoelectric sensors
2. Force sensing resistor
3. Antistatic foam
4. Force sensing by motor current monitoring
5. Strain gauge force / torque sensors
Piezoelectric sensors:
Piezoelectric material compresses if exposed to a voltage and produces a voltage if
compressed. This was used in devices such as the phonograph to create a voltage from the
variable pressure caused by the grooves in the record. Similarly, a piece of piezoelectric can be
used to measure pressures, or forces in robotics. The analogue output voltage must be
conditioned and amplified for use.

Figure 3.38 the piezoelectric sensor (LDT1-028K)


The LDT1-028K is a multi-purpose, piezoelectric sensor for detecting physical
phenomena such as vibration or impact. The piezo film element is laminate to a sheet of
polyester (Mylar), and produces a useable electrical signal output when forces are applied to
the sensing area. The dual wire lead attached to the sensor allows a circuit or monitoring device
to process the signal.

ME424U3ChLS201516 41
Force sensing resistor:
The force sensing resistor (FSR) is a polymer thick-film device that exhibits a
decreasing resistance with increasing force applied perpendicular to its surface. In one
particular model, the resistance changes from about 500 kΩ to about 1 kΩ for forces of 10 to
10000 gm. Figure shows a typical force sensing resistor.

Figure 3.39 a typical force sensing resistor (FSR)

Antistatic Foam:
The antistatic foam used for transporting IC chips is conductive and its resistance
changes due to an applied force. It can function as a crude and simple, yet inexpensive, force
and touch sensor. To use a piece of antistatic foam, insert a pair of wires into two sides of it and
measure the voltage or resistance across it. Figure 3.40 shows the usage of antistatic foam for
force measurement.

Figure 3.40 use of antistatic foam for force measurement

ME424U3ChLS201516 42
Force sensing by motor current monitoring:
It is possible to use motor current as a measure of the grasping force. If we assume that
the gripper itself is servo controlled and that the actuator is a servomotor, the armature current
(monitored differentially across a small resistance placed in series with one of the motor leads)
is proportional to the torque generated by the motor.
= ----------- (1)

Where, Ia is current, T is total torque generated, KT is motor torque constant.


Through the action of a mechanical rotary-to-linear motion converter (e.g., a rack and
pinion), this torque is converted into a force. With a pinion gear of radius R, the transmitted
force F is given by:
×
= ------------- (2)

Where, η is the efficiency of the rack and pinion assembly (usually on the order of 90%, i.e.,
η=0.9).
Substituting equation (1) in (2) we get,

From the above equation, it can be seen that when the robot is commanded to grasp a part,
the motor current can be increased to a value that produces sufficient force to prevent slipping
but is small enough to avoid crushing. These maximum and minimum values of I a would have
to be determined offline for every part to be handled and the information stored in the robot’s
controller.
Another factor here is that the actual force applied to an object by a gripper would be
somewhat smaller than that calculated by the above equation because of friction and gravity
loads. That is some motor torque (current) would be required to overcome the friction in the
various mechanical linkages used in a servo controlled gripper.

Another problem with this force controlling technique is that the temperature sensitivity of the
both the servo motor torque constant KT and its resistance Ra.

ME424U3ChLS201516 43
Strain gauge Force Sensors:
One of the simplest methods of sensing a force (pressure) exerted on an object is to detect
the deflection of the fingers of the robotic gripper in response to such an applied force. The
strain gauge provides a convenient and accurate means of doing this.
The principle of strain gauge is that a mechanical deformation produces a change in
resistance of the gage, which can be related to the applied force. Consider a simple strain gage
(figure 3.41) consisting of a plastic body (or some other flexible, non conducting base) whose
top surface is coated with a thin layer of a conducting material (e.g., aluminum or copper).

Figure 3.41 A simple strain gauge (a) un-deformed device-gage length L=1.5 in (b) concave up
deflection of an object L<1.5 in, and Rg decreases (c) concave down deflection of an object,
L>1.5 in, Rg increases

ME424U3ChLS201516 44
If the conductive coating is assumed to have a uniform cross sectional area A, then the
resistance of the device is given by

Where σ is the conductivity of the conducting material and L is the length of the gage.
When glued to an object, any deformation will cause the gage to bend either concave up or
down. Since the conducing material thickness is assumed to be small, this bending produces
little change in A. However, the gage length is either reduced or lengthened. It is seen from
above equation that such action causes a corresponding decrease or increase in the gage
resistance.
There are four basic types of strain gages:
1. Un-bonded wire
2. Bonded metal foil
3. Thin film
4. Semiconductor (Bonded bar & Diffused type semi conductor strain gauge)

Un-bonded wire gage usually consists of an extremely fine platinum tungsten wire.

In the case of metal foil gage, a thin copper or aluminum alloy foil is glued to either a non-
conducting base (bonded) or directly to the object under study (non-bonded).

With the thin film gages, vacuum or sputter techniques are used to deposit resistors onto
heat treated substrates. As such these devices represent the “state of the art” in strain gages.

The most sensitive of the four types is the semiconductor gage and is divided in to two sub
classes. One is made of silicon elements bonded to a cantilever beam or diaphragm and is
called a bonded bar semiconductor strain gage. The other one uses photolithographic and
diffusion techniques from integrated-circuit technology to produce a “diffused type
semiconductor strain gage”.

ME424U3ChLS201516 45
Regardless of what type is employed, some electronic circuit must be used to sense the change
in gage resistance and produce a voltage output as a result. Such a circuit is shown in figure
3.42.

Figure 3.42 strain gage bridge circuit

R1 is approximately equal to Rg, when the gage is not deformed. Rb is used to balance the
bridge before any deformation occurs. VAB or VOUT can be related to the applied force.
Here the gage is one arm of a wheat stone bridge (Rg). Another arm is adjusted so that its
resistance R1 is approximately equal to that of the unstressed gage. A 100 kΩ fixed resistor
R2and a 1MΩ balance pot Rb complete the circuit. If the bridge is initially balanced, so that
=
The output voltage VAB will be zero. Any change in Rg caused by a deformation of the gage
will unbalance the bridge and cause VAB to have a nonzero value. By applying a known force to
the object on to which this sensing device has been placed, the system can be calibrated. (i.e.,
the value of VAB can be related directly to force)

Applications of Strain gages:


Strain gages can be used to produce a robotic force sensing element. For example as
shown in figure 3.43, these can be placed on the back of the fingers of a parallel jaws gripper.
Then, as the fingers begin to grasp an object, the resultant deflection will be monitored by the
gages. If the deflection versus force characteristic for the gripper material and structure is
known, the grasping force can be related to the output voltage.

ME424U3ChLS201516 46
Figure 3.43 A robotic force sensor can be made by placing strain gages on the outer
surfaces of the fingers of a parallel jaws gripper (these gages monitor the finger
deflection when an object is grasped)

The principal approaches for force and torque sensing in robotics are joint and wrist sensing.
A joint sensor measures the Cartesian components of force and torque acting on a robot joint
and adds them vectorially. Wrist sensors are mounted between the tip of a robot arm and the
end-effector. They consist of strain gauges that measure the deflection of the mechanical
structure due to external forces.

WRIST SENSOR
Wrist sensors are small, sensitive, and light in weight and relatively compact in design on
the order of 10 cm in total diameter and 3 cm in thickness, with a dynamic range of up to 200
lb. In order to reduce hysteresis and increase the accuracy in measurement, the hardware is
generally constructed from one solid piece of metal, typically aluminum.

ME424U3ChLS201516 47
Figure 3.44 wrist force sensor
Figure 3.44 shows a typical wrist sensor, uses eight pairs of semiconductor strain gauges
mounted on four deflection bars- one gauge on each side of a deflection bar. The gauges on the
opposite open ends of the deflection bars are wired differentially to a potentiometer circuit
whose output voltage proportional to the force component normal to the plane of the strain
gauge. The differential connection of the strain gauges provides automatic compensation for
variations in temperature.
Since the eight pairs of strain gauges are oriented normal to the x, y, and z axes of the force
coordinate frame, the three components of force F and three components of moment M can be
determined by properly adding and subtracting the output voltages respectively. This can be
done by pre multiplying the sensor reading by a sensor calibration matrix given below.
Most wrist force sensors function as transducers for transforming forces and moments
exerted at the hand in to measurable deflections or displacements at the wrist. It is important
that the wrist motions generated by the force sensor do not affect the positioning accuracy of
the manipulator. The required performance specifications of these sensors are high stiffness,
compact design, linearity, low hysteresis and internal friction.
Assuming that, the coupling effects between the gauges are negligible, that the wrist force
sensor is operating within the elastic range of its material, and that the strain gauges produce

ME424U3ChLS201516 48
readings which vary linearly with respect to changes in their elongation. Then the sensor
produces eight raw readings which can be resolved by computer software, using a simple force-
torque balance technique, into three orthogonal force and torque components with reference to
the force sensor coordinate frame.
Mathematically,
=
Where, F = (force, moments) T = (Fx, Fy, Fz, Mx, My, Mz) T
W= raw readings= (w1, w2, w3… w8) T
r11 . . . r18
RF= . . . . .
r61 . . . r68
With reference to figure 3.44,

0 0 r13 0 0 0 r17 0
r21 0 0 0 r25 0 0 0
0 r32 0 r34 0 r36 0 r38
RF= 0 0 0 r44 0 0 0 r48
0 r52 0 0 0 r56 0 0
r61 0 r63 0 r65 0 r67 0

Figure 3.45Appication of wrist sensor on a robot

ME424U3ChLS201516 49
Review Questions
1. Discuss in detail the classification of robotic sensors.
2. With a neat sketch explain the principle and operation of scanning laser proximity sensor.
3. What is the principle of synchro? Explain the applications of synchro.
4. With a neat sketch explain the operation of resolver.
5. Differentiate position measurement by absolute and incremental encoders. List out their
features.
6. How potentiometric sensor is useful for robotic manipulator?
7. What is slip sensor? Explain its importance in terms of robot grasping a work part.
8. What is a tactile sensor? Explain its importance in robotic applications.
9. Explain the importance of force and torque sensing in robots.

Problems

1. What will be the change in resistance of an electrical resistance strain gauge with a gauge
factor of 2.1 and resistance 100Ω if it is used to measure a strain of 0.1%?
2. What is the non-linearity error, as a percentage of full-scale reading, produced when a 1kΩ
potentiometer has a load of 10 kΩ and is at one-third of its maximum displacement?

ME424U3ChLS201516 50

You might also like