Robotics

Prof. Dilip Kumar Pratihar

Lecture – 31
Sensors

We are going to discuss another topic, that is, topic 6, that is on Sensors. Now, let us see
like how to design and develop the sensors, how to use the sensors, what are the different
types of sensors used and how can we collect information with the help of sensors?

(Refer Slide Time: 00:34)

Now, we human-beings, we use different types of sensors like we have got the eyes, ears,
nose, skin. In fact, we use multiple sensors to collect information of the environment.
And, the data collected with the help of this multiple sensors are actually processed in
our brain and with the help of this particular processing, we can collect information of
this particular environment.

Similarly, if you want to make robot intelligent, we should put a few sensors and these
sensors will help the robot to collect information. Now, here, let me define that the
sensor is nothing, but a transducer and we generally use sensor to take some
measurements of physical parameter or physical variable and here this sensor; if you
want to use as a measuring device; so, definitely there must be some calibration.
(Refer Slide Time: 01:41)

And, by calibration actually we mean, it is actually the comparison with some known
data. Now, through comparison with the known data, we will be able to calibrate a
particular measuring device or a particular sensor.

Let me take a very simple example, supposing that I will have to draw a straight line of
say 10 millimeter. So, starting from here; so I am going to draw supposing that this is my
10 millimeter straight line, ok. Now, if I am told that can you not draw one another
straight line which is 20 millimeter long. So, what I will do is if this is 10; so my eyes are
going to measure with the previous one and might be this is 20. So, this will be 20
millimeter.

That means, my eyes are following some sort of calibration; if this is 10 millimeter; so,
this will become the 20 millimeter just double of that. So, our eyes while taking this
particular information; it is following some calibration; it is following some calibration
scale and the same is true for any such sensor. Now, you might be knowing that we use
different types of sensors, we use different sensors to take some measurement, for
example, to measure the joint torques, we use sensor, to measure force, we can use some
sensor, but will have to calibrate.

Now, this calibration is a must for any such sensor, now here actually if you want to
make it intelligent; as I have already discussed that sensors are to be used or cameras are
to be used to collect information with the help of to collect information of the
environment and then only we can do some processing to take some decision in a very
intelligent way; so, this particular calibration is a must for any such sensor. Now, if you
see the literature, we have got different types of sensors for example, say, if you classify
the sensors in this way, we can classify like internal sensor and external sensor.

Now, these internal sensors are nothing, but the sensors, which are used to operate the
drive units. For example, we have got some position sensors, then we have got the
velocity sensor, acceleration sensors, then force or the moment sensors; these are all
internal sensors. And, on the other hand, we have got a few other sensors, which are used
to collect information of the environment and those are known as the external sensors.
For example, we can use some sort of proximity sensor, acoustic sensor, then comes your
visual sensor, temperature sensor; these are all external sensors.

Now, here, if you see in our human body, we have got a few internal as well as a few
external sensors. For example, say, whenever we try to collect information of the
environment; we try to use our eyes. So, with the help of the eyes, we collect information
of the environment, but supposing that we are getting some pain in the muscle of leg;
now how can we feel that there is some pain? So, to feel that particular pain in the
muscle; we use some other types of sensors and those are known as the internal sensors.
So, in our human body, we use internal as well as external sensors, the same is true in
robots.

In robots also, we use a few internal sensors, we use a few external sensor. Now, the
working principle of the different internal and external sensor used in robots, I am just
going to discuss one after another, in details.
(Refer Slide Time: 06:04)

Now if you see the literature, the sensors are also classified in different ways for
example, the sensors could be named as contact sensor or the non-contact sensors. So, by
contact sensors, we mean that there is a physical contact between the sensor and this
particular object, whose distance I am going to measure. And, if there is no such contact,
the physical contact between the sensor and the object it is called the non-contact sensor.

Now, this contact sensor can be further classified into two subgroups, now, one is called
the touch sensor or tactile sensor or the binary sensor. So, we have got the touch sensor
or tactile sensor or this binary sensor, it is almost similar to our skin, skin is nothing, but
our touch sensor. So, in robots also we use some touch sensor or the binary sensor for
example, say the micro-switch or the limit switch, which are generally used in robots are
nothing, but the tactile sensor or touch sensor.

Now, with the help of this touch sensor; so, it is simply going to tell that the robotic
finger has touched a particular object, but it is not going to measure the force required to
grip that particular object or how much is the force required to or how much is the torque
required to manipulate that particular object.

So, it only indicates whether the contact has been made or not now let me take a very
simple example. Now in all the water tanks or the oil tanks we use a valve that is called
the float valve. Now this particular float valve, what is the function of the float valve? If
this is the tank, the moment the water height reaches a particular level the peaks level,
this particular float valve will be activated and it is going to indicate that we should stop
the pump. And, the pump will be stopped and the water supply to this particular water
pump will be stopped. So, this indicates the highest limit, the highest permissible limit
for this particular the water. And, that is nothing, but this float valve is an example of the
limit switch or either the touch sensor or the tactile sensor.

Now, this is also known as the binary sensor for example, say the micro switch or the
limit switch which is generally used in robotic hand. I am just going to take one example,
the next slide I will show you that we can use some sort of micro-switch or the limit
switch along with your robotic hand.

Now, this is also called the binary sensor because it generates 1s and 0s. The moment it
touches; it will generate, it will generate this particular 1 and otherwise it will generate
this 1; so, it is going to generate 1 or 0s ok. So, this is known as binary because
sometimes it generates 1, sometimes it generates 0; if there is a contact it will generate 1,
if there is no contact it will generate 0. So, it is some sort of 1 0 0 1 something like this
and that is why, this is also known as the binary sensor.

Now, I am just going to concentrate on the force sensor or this analogue sensor. Now, as
I told that with the help of this force sensor or this analogue sensor, we are just going to
actually measure the force or the torque. So, this particular portion torque which is
required to grip this particular the object and generally, we use some sort of the force
sensor or the analogue sensor.

And, here in force sensor or analog sensor, we use some sort of strain gauges. So, I am
just going to discuss the working principle of this particular strain gauge in details. And,
as I told that we have got a few non-contact sensors, for example, say we have got the
proximity sensor, the range sensor, the visual sensor, acoustic sensor.

So, these are all non-contact sensors. So, I am just going to discuss, in details, the
working principle of these sensors generally used in robots.
(Refer Slide Time: 10:55)

Now, before I go for the discussion of working principle of different sensors; now let me
concentrate a little bit on the different characteristics of sensors. Like, if you want to
prepare the specification of sensor; what are the information to be provided and what are
the numerical values to be provided?.

For example, say the range, response, accuracy, sensitivity, repeatability, resolution all
such things, we will have to mention. While preparing the specification of the robot the
similar type of information, we also provided. And, here actually or the range for the
sensor; that means, what is the maximum and the minimum value that can be measured
with the help of this particular sensor that has to be mentioned.

Then, comes your the response, the response should be as quickly as possible, then
accuracy is nothing, but the deviation from the exact quantity. So, that we will have to
mention, sensitivity we know by definition sensitivity is nothing, but the change in
output to the change in input; so, that is nothing, but your sensitivity. So, this particular
sensitivity, we will have to mention, how much sensitivity you need and if this particular
sensor is having constant sensitivity; now then it is called the linear. So, the sensor is
called a linear, if it is having the constant sensitivity.

So, by linearity, we mean constant sensitivity; then comes your repeatability. Now, by
repeatability, we know supposing that with the help of the same sensor, I am just going
to measure the same thing for say 10 times or 20 times. Now, the same thing, if I
measure 10 times or 20 times; there is no guarantee that all 20 times will be getting
exactly the same numerical value.

Now, this particular deviation from reading to reading is nothing, but the repeatability.
Now, while preparing the specification of this particular sensor; we will have to mention
how much is the repeatability we want? And, then comes your the resolution is nothing,
but the least count. So, this particular list count for the measuring device or this
particular sensor; that we will have to know. Now, let me take a very simple example
now, if I take one very simple example for this particular resolution, it will be clear.
Supposing that I am using one sensor and in that particular sensor, I am using some
electrical signal to generate some angular displacement.

Now, electrical signal cannot be a fraction. So, it could be 1, 2, 3 something like that;
corresponding to one electrical impulse, how much is the angular displacement it can
generate? That is nothing, but the least count or resolution, the same is true for this
sensor, ok. So, these are the information which are to be provided to prepare the
specification of this particular sensor.

(Refer Slide Time: 14:35)

Now, I am just going to discuss the working principle of a few sensors one after another.
Now, this touch sensor, I have already discussed; these are use just to indicate whether
the contact has been made or not. And, generally, we do not use this sensor to determine
how much is the contact force? The examples are micro-switch, limit switch and all such
things.

(Refer Slide Time: 15:01)

Now, here, I am just going to take one typical example of a micro-switch, which is
nothing, but a touch sensor used in robot gripper. Supposing that this is a very simple
gripper having two fingers and here we are just going to put some micro switch or the
limit switch. Now, with the help of this micro-switch or the limit switch actually, it is
going to indicate whether the contact has been made between the object; supposing that I
have got the object here, whether the contact has been made between the object and this
particular the robotic finger.

So, to serve that type of purpose, we use your the micro-switch or the limit switch and
that is nothing, but the touch sensor, the same is true, we have got the skin. So, with the
help of this particular skin, we can touch, we can feel the presence of an object, even if
we are not using eyes; we can feel the shape and size or the more or less the structure of
that particular object, even if we do not see with the help of our eyes.

Because we can use our skin and skin is nothing, but the touch sensor and with the help
of this touch sensor; in fact, we can find out, what should be the possible shape and size
of this particular object, which I am going to grip.
(Refer Slide Time: 16:31)

Now, I am just going to discuss one sensor one position sensor which is very frequently
used in robotics or we generally use in school level, college level some laboratory classes
also. This is your the potentiometer the potentiometer is actually why as I told very well
known position sensor. And the potentiometer could be either the linear potentiometer or
it could be angular potentiometer.

Now, with the help of linear potentiometer, we can measure the linear displacement, that
is, d and with the help of angular potentiometer, we can measure the angular
displacement, that is, nothing, but θ . Now, the working principle of this particular
potentiometer is very simple; for example, say so here we have got the source for the
voltage that is input voltage V_in for example, say we have got the battery or V is
connected to some power, ok.

Now, what we can do is supposing that we have got the battery here so, I know the input
voltage. So, what I do is, we know the total resistance of this particular wire; so, this is
actually nothing, but the wire and it has got a special type of winding. So, capital R is
nothing, but the total resistance and the reference here. So, this is nothing, but the
reference that means, we are going to measure the angular displacement with respect to
this particular reference, ok.

Now, here actually, what I do is, we are going to measure this angular displacement with
the help of this pointer or the wiper. So, with respect to the reference; so it is going to
generate some angular displacement ok; how to measure this? To measure this, the
method which we follow is very simple; so, here with respect to this particular reference.
So, I have got the wiper and it has got some displacement and we measure with the help
of a voltmeter, how much the output voltage? So, this one point is connected to the wiper
and another is your the grounded. And, you can find out, we can measure with the help
of voltmeter, how much is the output voltage?

So, we know the total resistance of this particular wire, we know how much is the input
voltage, we can measure, how much is the output voltage with the help of the voltmeter
or multi-meter. And, if you have measured this particular output voltage; now
approximately I can find out what should be this angular displacement. Now, how to find
out? It is very simple, because if I know this input voltage and if I know the resistance.

Vinput
So, the current is nothing, but your and the same current will also flow here and
R
Vout
that is nothing, but is your and small r is nothing, but the resistance of this winding
r
up to starting from the reference; up to the end of this particular pointer or the wiper. So,
starting from here up to this; so this is the resistance small r, so from here see the V_in is
known, capital R is known, V_out can be calculated.

So, r can we determined and if I know the value of r and if I know the nature of winding
of this particular wire, the electrical wire. I can find out approximately like what should
be this particular angular displacement, that is, θ . So, θ can be measured with the help
of this particular angular potentiometer. This is the working principle of angular
potentiometer, it is very simple and all of us have used.
(Refer Slide Time: 20:54)

So, this is the working principle of your angular potentiometer, but this angular
potentiometer has got one demerit or one drawback; all of us we know that the resistance
of wire depends on actually the temperature.

The moment we pass some current through electric wire; so, due to the heating effect of
the current, that is, i 2 r effect. So, what will happen? There will be some heat generated
in that particular electric wire and its temperature is going to increase and as temperature
increases; so the resistance of the wire is going to change and if resistance changes, and
you will not be getting very accurate measurement with the help of this angular
potentiometer.

And, that is why, if we use this particular angular potentiometer at a stretch for a long
time; so, initially we may get some accurate results, accurate measurements. But, with
time, after might be half an hour or one hour, there is a possibility, we will be getting
some erroneous results with the help of this angular potentiometer. So, this is actually the
drawback of this angular potentiometer, but its working principle is very simple and this
is, in fact, one of the most popular position sensor used nowadays.
(Refer Slide Time: 22:28)

Now, then comes your another position sensor that is called the optical encoder. And,
this is also very popular and if you see we have got two types of optical encoder. One is
called the absolute optical encoder and we have got the incremental optical encoder. And
in robotics actually very frequently this type of optical encoder is used as feedback
device. For example like, if it is servo-controlled robot; so, there must be a provision of
feedback device and there must be a provision to measure the angular displacement. So,
what you can do is; we can take the help of, so this type of optical encoder as a feedback
device.

Now, let us see, how does it work? Now, this optical encoder, now, let me first try to
explain the principle of this absolute optical encoder first. Now, this absolute optical
encoder consists of a number of concentric rings placed one after another. Now,
supposing that say this is the output shaft, this is the output shaft of this particular motor
I have got the electric motor here.

So, this output shaft is rotated; now I want to measure what should be the angular
displacement or what is the rotation of this particular shaft? What I do is here I put this
particular absolute optical encoder and absolute optical encoder is nothing, but a
collection of a few concentric rings placed one after another and what I do is. So, here
we have got the concentric rings and on this particular concentric rings; there will be the
marking zone; that means, there will be dark zone and the light zone, and through the
dark zone, the light will not pass and for through this particular light zone, the light will
pass now on.

So, here, we have got the optical encoder; so, as the shaft rotates the optical encoder
mounted on it is also rotating. Now, on one side, I have got the photo source, other side,
we have got the photo detector. The moment during the rotation; this particular disc, the
circular disc or the rotating disc; if the light zone comes in front of the light source, the
light will pass and it is going to activate that particular photo-detector. The same thing I
am just going to discuss in more details with the help of this particular the sketch.

Now as I told that it is consisting of a large number of concentric rings. Now here for
simplicity I am just going to consider; so, there are only 4 concentric rings. Now
supposing that this is nothing, but the diameter; this is the diameter of the shaft whose
angular displacement or rotation I am going to measure.

Now, here surrounding this we consider say one concentric ring, another concentric ring,
another concentric ring. So, I am considering 4 concentric rings here now you
concentrate on the first one that is this particular concentric rings. Now, here what I do
is; so this part is made black, this part is made black on the first concentric ring. So, this
is made black this part is made black, and this is your the white portion through which
the light will pass. Next, we concentrate on the second concentric rings on the second
concentric rings starting from here, up to this is made black; so, this part is made black.

So, no light will , then it is white once again, this part is made black and then, there is a
white portion, then you concentrate on the third one. So, here, this is the black part and
this is the white part, then this is the black part, then comes a white part, then comes this
is the black part, this is the white part, this is the black part, then comes the white part
and so on, and on the outermost ring, we have got the black portion here.

So, this is the black part, white part, black part, then white, black, white, black and so on.
So, this type of marking we have and here we are considering only 4 concentric rings.
The outermost ring is going to indicate actually 20 . And, this particular thing is going to
indicate 21 , this is 22 and this indicates 23 and so on.

Now, if you see the other view for example, say in this particular view, supposing that
this particular shaft it is not drawn here properly. So, this particular shaft supposing that
this is actually the diameter of the shaft and on which, I have got this concentric rings
mounted one after another ok; that means, this is going to indicate 20 ; that is the outer
most and this is going to indicate 23 .

So, here this is another view; so on this side, we have got the light source and this
particular thing is rotating. So, this particular thing is rotating; this is mounted on the
shaft, the shaft is rotating, optical encoder is also rotating and the light source is put on;
the moment the dark zone is coming. So, here, there will no signal, the moment I will be
getting the light zone then only there will be light will pass through this and it is going to
activate that particular photo-detector.

So, depending on the relative position of the dark zone and the light zone; so, sometimes
light will pass, sometimes it will not pass accordingly it will be generating some 1s and 0
sort of thing, ok. For example, say here, this is the reference line; initially the reference
line is here, ok. Now, the reference line is fixed and this particular optical encoder is
rotating; now what I do is, here on the screen, I cannot rotate. So, what I can do is; I am
just considering as if this optical encoder is fixed and I am rotating this reference in the
reverse direction. If I just the rotate the reference in the reverse direction.

The moment my reference is here, the moment my reference is here, truly speaking, the
optical encoder is rotating reference is kept fixed. But, here, this is almost equivalent of
the situation, my optical encoder is fixed because I cannot rotate it here on the screen; so,
I am rotating the reference in the opposite direction. Now, supposing that the reference is
here and if it is here, then this is the dark zone, dark zone, dark zone and dark zone..

So, it is going to generate 4 such 0s, say this is 20 , 21 , 22 , 23 the moment this particular
reference comes here supposing that it is here. So, through the outer-most, the light will
pass; this outer most corresponds to a 20 . So, here it is going to generate 1, but
corresponding to 21 ; there will be 0, there will be 0, there will be 0; so 0, 0.

Similarly, the moment we consider that this particular reference is here. So, what will
happen? The light will pass through all four; so, through here light will pass, light will
pass through this, light will pass through this, light will pass through this. So, it is going
to generate actually four such 1s ok; so if it generates four 0s; its decoded value will be
equal to 0. Because 0 multiplied by 20 ; so decoded plus 0 plus 0 plus 0 it will be 0.
And, the decoded value for this; so, 1 multiplied by 20 that is equals to 1, plus 0
multiplied by this equals to 0, 0 and 0; its decoded value will be 1. And, corresponding to
this 1 1 1 1 the decoded value will be your something like this
1× 20 + 1× 21 + 1× 22 + 1× 23 .

So, 8 plus 4 that is your 12 plus 2; 14 plus 1; so, 15. So, its decoded value will be 15 ok.
So, corresponding to rotation of this particular your optical encoder and depending on
the position of angular displacement with respect to the fixed reference; I will be getting
some binary. This binary will be generated here and I can just do the decoding and I will
be getting, the decoded value corresponding to that particular rotation.

(Refer Slide Time: 32:47)

And, whatever I discuss, the same thing I have written it here. So, corresponding to this 0
0 0 0, I will be getting 0 and this is the way actually, I will be getting this particular the
decoded value. Now, supposing that I have got some decoded value; now if I use like
four concentric rings, if I use four concentric rings then your, how many divisions? We
are getting only 16 divisions.

So, 16 divisions is nothing, but your 24 . So, 24 is actually your 16, ok; that means,
corresponding to the whole rotation for this particular shaft that is nothing, but 360
degree corresponding to one rotation. So, this particular 360 degree, I am just going to
divide equally into 16 parts; that means, your; so, this 360 divided by 16 will be the
resolution of this particular optical encoder if I use only 4 concentric rings.
Similarly, if I use n number of concentric rings, then it will have the resolutions like 1
part in 2n . So, this is nothing, but the resolution of this particular optical encoder for
example, say if I take n equals to say 10; then the resolution will be 1 divided by actually
210 , that is nothing, but approximately that is equal to 1024 and approximately that is
equal to 1 divided by say 1000, ok.

That means your 360 degrees rotation for one complete revolution will be divided into
1000 equal parts and that is nothing, but 0.36 degree. Now, this particular 0.36 degree
will be the resolution of the optical encoder, if we use actually 10 concentric rings. Now,
this is the way actually with the help of this absolute optical encoder; we can measure
how much is the angular displacement of a particular shaft, which is rotating.

This is the working principle of this absolute optical encoder which is very frequently
used in robots as a feedback device.

Thank you.