You are on page 1of 26



ANAND RAJ.G.R. (13402014)




October 2016




This is to certify that the Seminar report titled MOLECULAR COMMUNICATION

is a bonafide record of the seminar conducted by ANAND RAJ.G.R., under our guidance
and supervision towards partial fulfilment of the requirements for the award of Bachelor of
Technology Degree in Electronics and Communication Engineering from the University of
Keraladuring the year 2016.

Seminar guide

Seminar Coordinator

Seminar Coordinator


Mr.Nelwin Raj N R

Smt.Subha V

Assistant Professor

Assistant Professor

Assistant Professor

Dept. of ECE, SCTCE

Dept. of ECE, SCTCE

Dept. of ECE, SCTCE

Head of the Dept. of ECE

Prof(Dr).Sheeja M K


I would like to express my sincere gratitude to the head of the institution

Dr.JayasudhaJSforproviding me with the opportunity to complete this seminar. I would also
like to thank Dr.Sheeja M K, Head of the department of ECE for her valuable remarks for
presenting this seminarsuccessfully. Also I would like to express my boundless gratitude to
my staff advisor andseminar coordinator Sri. NelwinRaj N R and my guide Smt. Sandhya L
for their valuablesuggestions which were imperative in the successful undertaking of my
seminar.Last but not least I would like to thank my classmates without their patience and
support thisseminar would not have been successful. I would also extend my heartfelt thanks
to my familyand well-wishers.


This topic involves the first modular, and programmable platform capable of transmitting a
text message usingchemical signalling

a method also known as molecular

communication.This form of communication is attractive forapplications where conventional

wireless systems perform poorly, from nanotechnology to urban health monitoring. By
providing a simple and inexpensive means of performing experiments,this system fills an
important gap in the molecular communication literature, where much current work is done in
simulationwith simplified system models. A key finding in this topic is that these systems are
often nonlinear in practice, whereascurrent simulations and analysis often assumes that the
system is linear. However, despite thenonlinearity, reliable communication is still possible.
Furthermore, this work motivates future studies on more realisticmodelling, analysis, and
design of theoretical models and algorithms for these systems.


1.1What is molecular communication? ................ 7
1.2 History ............................................................. 7

2.Materials and methods.................................. .........8

2.1 Design Criteria............................................................ 8
2.2 Hardware layout


2.2.1 Transmitter ................................................................ 9

2.2.2 Signalling chemical ...................................................11
2.2.3 Receiver .....................................................................11
2.2.4 Propagation channel.................................................. 12
2.2.5 Signal modulation and demodulation.......................... 14
2.2.6 Communication protocol design.................................15

3.Results and discussion ...............................................19

3.1 Noise reduction ................................................................. 19
3.2 Transmission rate ...............................................................19
3.3 Detection and demodulation algorithm .............................. 21

4. Conclusion and future work ....................................23

Reference............................................................................ 24

List of figures:
Figure 1: Block diagram of typical communication system .................. 9
Figure 2: Transmitter components ........................................................11
Figure 3: Receiver components .............................................................12
Figure 4: Diffusion based versus flow based propagation .....................14
Figure 5: Flow chart representation that controls transmitter ................16
Figure 6:Flow chart representation that controls receiver .....................18

List of Tables:
Table 1: Different transmittion ratesand reliability ...............................20

1.1What is Molecular Communication?
This topic presents the worlds firstmacroscale molecular communication link to
reliably transmit acontinuous data stream. The system modulates alcohol molecules,which
are then diffused via ambient and induced air currents tocarry information to a receiver.
The communication distance isseveral meters and the propagation channel consists of both
free space and tunnel environments[2]. The goalis to show that molecules can be used as
an alternative toelectromagnetic (EM) waves in challenging environments whereEM
waves do not perform well[21-24].

1.2 History
The transfer ofdata using chemicals between organismshas been observed throughout
history. Most notably, CharlesDarwin observed the relationship between chemical signalsand
mating patterns in 1871. In the 20th century, the termpheromone was coined to mean a
















communications[26]. There have alsobeen a number of attempts at mimicking pheromonebasedcommunication[4]. More recently, this team has proposed anddemonstrated that
molecular communications over macroscaledistances is not only possible, but also reliable.
Thisdemonstration for the first time

hasshown the worlds firstmacroscale molecular

communication link.

2.Materials and Methods:

2.1 Design Criteria:
A simple, robust, cost-effectivecommunication system uses chemical signals for
carryinginformation from a transmitter to a receiver. To test this system, a short text message
is sent.

To design and develop this system, the following criteria is followed[1]:

The end product must be inexpensive to build. This wouldmake the platform readily
available for many differentresearch and development projects with limited amount

The designed system must be simple and robust, much thesame as the telegraph, the
ancestor of modern telecommunicationsystems.

The developed system must be easily modifiable andprogrammable. Again this is an

important criterion for futureexpansions and adoption to different applications.

Any communication system can be broken down into threemajor parts: the transmitter, the
receiver, and the channel.Figure1shows block diagram representation of these three
modulesand their submodules.

Figure 1.Block diagram of a typical communication system.

2.2 Hardware layout

The transmitter takes an input text message from a user. It thenconverts the text
message into a sequence of binary bits andmodulates them on a chemical signal for
propagation in thechannel.

To convert the text message to a binary sequence, we use theInternational Telegraph

Alphabet No. 2 (ITA2) standard [27],where every letter is represented using five bits. For
example, theletter E is represented by a five bit sequence 10000. Forsimplicity, this
work does not use any error-correcting code.
Therefore, the five bit encoded letters are passed to the modulator block of the
transmitter for modulation and transmission to thechannel.

Transmitter Design:
On the transmission side, the hardware consists of:

A user inter-face for text entry,


a microcontroller that convertsthe input text into binary sequences and then
modulates thesequence on chemical signals,


a reservoir of chemicals,

iv) a chemical release mechanism.

User interface:
For text entry, the 1662 character LCD Shield Kit from Adafruit is used. The LCDis
an add-on module for the Arduino microcontroller board, whichalso has six push buttons. A
program is written in the Arduinomicrocontroller which employs the LCD and its buttons for
textentry by the user[.

To control all transmission operations, theArduino Uno open-source electronics
prototyping platform is used, whichis an ATmega328 based microcontroller board.

Chemical Reservoir and Chemical Release Mechanism:

To modulate the channel symbols into chemical signals, an electronic spray called
DuroBlast made by Durotech Industries is used.The DuroBlast electronic spray has a battery
operated electricalpump that can spray a wide variety of liquid chemicals that can bestored
inside its container. A custom electrical switchboard that can be used to control the spray
from the Arduino microcontroller board was designed. By programming the Arduino
microcontrollerboard, any type of modulation can be implemented throughcontrolled set of
Figure 2 shows the transmitter setup withall of its subcomponents.


Figure 2.The transmitter components.

2.2.2 Signalling Chemical:

To achieve the design criteria, the receivers sensor must besensitive, widely
available, and inexpensive. Moreover, it must beable to detect a volatile, widely available,
and inexpensivesignaling chemical that is safe at the low concentrations. Therefore,
isopropyl alcohol (rubbing alcohol) was chosen asthe signaling chemical.

To design the receiver, a sensor is required that is capable of detecting achemical
signal. The data from the sensor is processedby the demodulation and detection algorithms,
and finallydecoded into text.

Receiver Design:
On the receiver side, the hardware consists of:

a chemical sensor,


a microcontroller that demodulates and decodes the signal, and



a computer for displaying and visualizingresults.

Chemical Sensor:
Three different candidate sensors fordemodulation and detection at the receiver: MQ3, MQ303A,MR513 alcohol sensors, all of which are manufactured by HenanHanwei
Electronics Co. Ltd. of China. All three sensors use ametal oxide semiconductor detection
layer for detecting thealcohol[27], but each has a different sensitivity, power and
operationcircuit diagrams. Besides isopropyl alcohol, the sensors can detectother types of
alcohol such as ethanol. However, in this workonly use isopropyl is used. All three sensors
were implemented on a custommadePCB board as shown in Figure 3.

Figure 3.The receiver components.

Microcontroller and display:

Again the Arduino Uno open-sourcemicrocontroller is used for programming and
controlling all the receiveroperations. The Arduino Uno board has a 10-bit analog to
digitalconverter that can be used to read the sensor data. Thedemodulation and detection
block and the source decoder blockcan then be programmed into the microcontroller, and
theresulting detected text message can be displayed on a computerscreen using serial port.


2.2.4 Propagation Channel

Two different propagation schemes were considered for the channel:
diffusion, and flow assisted propagation.

In the diffusion basedpropagation, after the initial spray the alcohol diffuses in the
airuntil it reaches the receiver. The diffusion propagation does not require external energy
(beyondthe energy required to release the chemical message)In flow based propagation, a
tabletopfan is used to guide the alcohol towards the receiver.

Flow assisted propagation:

It requires external power. We use twodifferent tabletop fans to generate flow:N
Honeywell 7 inch Personal Tech fan: This fan is aninexpensive bladed fan (approximately
$16 USD) with twodifferent fan speeds low and high.N Dyson AM01 10 inch bladeless fan:
The Dyson fan is muchmore expensive (approximately $250 USD), but can generatemore
laminar flows and many different wind speeds byadjusting an analog nub.When any of the
two fans are used, they are placed 30 cmbehind the spray.To measure the flow speeds
generated by each fan thePyle PMA82 digital anemometer is used. The maximum flow speed
ismeasured at distances of 10 cm, 50 cm, 100 cm, 150 cm, and200 cm from the front of the
spray (the fan is placed 30 cm behindthe spray).

The system response (the output of the sensor for asingle short spray) under both
propagation schemes (diffusion andflow assisted propagations) was compared. At short
distances (up to 1 meter), thediffusion based propagation scheme performs well because
thealcohol ejected from the spray reaches the sensors almostinstantaneously. However, if the
spray is placed further away,diffusion based propagation would not be practical because of
theextremely slow system response[5]. This effect can be seen in Figure 4,where the system
response to a very short sprayof 250ms induration is plotted for both diffusion based and flow
basedpropagations. The flow in this figure is generated usingHoneywell fan on the high
setting, and the spray is placed at adistance of 2 meters from the detection sensor.


Figure 4.Diffusion based propagation versus flow based propagation.

As can be seen, thesystem has a quick and distinct response when flow based
propagation is employed. Although the response is plotted for onlyone of the sensors (MQ-3
sensor), the same effect was observed forall the other sensors, as well as when the Dyson fan
is used in placeof the Honeywell fan. Therefore, for the molecular communication setup
flow based propagation is used.

2.2.5 Signal Modulation and Demodulation:

Because communication is performed through chemical signals,and a limited amount
of signalling chemical can be stored in acontainer at the receiver, the modulation and
demodulationscheme selected should minimize the amount of chemical used.The source
coding scheme for encoding text messages, presentedin previous sections, uses the least
amount of ones in the 5 bitsequence for characters that have a higher rate of occurrence inthe
English text. For example, letters E and T both have asingle one in their 5 bit


sequences. Therefore, the bit1 is modulated with a single spray and the bit 0 with no spray.
Thismodulation scheme, which is called on-off keying, effectivelyminimizes the amount of
chemical used for communicatingEnglish text.

At the receiver the demodulation is performed by measuring therate of change in
concentration. If during a single bitscommunication session the voltage reading from one of
thesensors is increasing (i.e. the concentration of the chemical signal isincreasing), then the
signal is demodulated as the bit 1. Similarly, ifthe voltage reading from one of the sensors is
decreasing (i.e. theconcentration of the chemical signal is decreasing) the signal
isdemodulated as the bit 0.

2.2.6 Communication Protocol Design

This section involves the communication protocol betweenthe transmitter and the
receiver, and its implementation. Indesigning the protocol the criteria that the protocol must
besimple, asynchronous[6] (i.e. no synchronizationis required betweenthe transmitter and the
receiver), and should work independent ofthe separation distance between the transmitter and
the receiver is used.
(i.e. it should not only work for a predefined fixed distancebetween the transmitter and the

Transmitter side:
At the transmitter, the output of the source encoder (i.e. the bitsequence representing
the text message) is concatenated with a twobit sequence 10 at the beginning and the null
characterrepresented by 00000 at the end. The initial 10 indicates startof a text
message and the null character indicates the end of thetext message. For example, if the text
message that is beingtransmitted is the letter A, the output of the source encoder isthe five
bit sequence 11000 (where the left most bit position isthe first bit position), and
transmission bit sequence is101100000000. The transmission bit sequence is then
modulatedusing the scheme discussed in the previous section, where 1 ismodulated with a
spray and 0 with no spray.
Figure 5 representsthe flowchart of the algorithm that runs at the transmitter, andsummarizes
this process.


Figure 5.Flowchart representation of the algorithm thatcontrols the transmitter.

Receiver side:
At the receiver, there are two states: the wait state, and thereception state. In the wait
state the receiver uses its sensor tocontinuously monitor the concentration of alcohol. If there
is asudden increase in the concentration of alcohol (i.e. suddenincrease in the sensors voltage
output), the receiver switches to thereception state. This sudden change is caused by the
initial 10bit sequence concatenated to the beginning of every text messagesent by the
transmitter. This sudden change can also be used as thereference time for synchronizing each
bit interval for all the bitsthat would follow. Therefore, no synchronization is
requiredbetween the transmitter and the receiver in advance. Anotherfactor that is taken into
account in this scheme is the propagationdelay[7]. Because the receiver is triggered into
reception state as soonas the leading bit 1 is detected, the time delay caused by
signalspropagation over the separation distance from the transmitter tothe receiver is
incorporated in the reference time[15-20].


Therefore, thecommunication protocol is independent of the distance betweenthe transmitter

and the receiver, and it would work even when thedistance is changed between
communication sessions.

After the receiver enters the reception state, it waits for two bitintervals until the
reception of the initial 10 bit sequence isfinished. The receiver then demodulates and
decodes the receivedsignal 5 bits at a time. During each 5 bit interval, the sourcedecoded
character is displayed to the computer screen using serialport connection. This process
continues until the null characterrepresented by all zero sequence 00000 is detected.
Because thenull character indicates the end of the text message, the receiverwill go back to
the wait state until another text message is sent bythe transmitter. Figure 6 summarizes the
algorithm that isimplemented at the receiver.


Figure 6.Flowchart representation of the algorithm thatcontrols the receiver


3.Results and Discussion:

To make this system more cost effective, the Honeywell fan was used for
demonstration despite the fact that theDyson fan can create better system response.The MQ-3
sensor was used because it provides the best system response and it has thesimplest circuitry
between all three sensors.

3.1 Noise Reduction:

Although the MQ-3 sensorresponse is less corruptedby noisecompared to the other
twosensors, there is still some noise present in the signal. To furtherreduce this noise, 20 ms
of sensor data is averaged to generate asingle sensor reading. Because the Arduino
samplingrate isobserved to be about 8.33 kHz, 20 ms of sensor data contains 167different
readings which are then averaged. Therefore, in the waitstate the receiver checks consecutive
20 ms of averaged sensorreadings, and triggers a change to the reception state if thedifference
between the current reading and the previous reading isgreater than 0.5 levels (because
Arduino has a 10 bit analog todigital converter the sensor reading would be an integer
between 0and 1023 representing 1024 different voltage levels, where 0represents 0 volts and
1023 represents 5 volts).

3.2 Transmission Rate:

An important communication parameter is the transmissionrate. One of the major
factors that affects reliable communicationat a given transmission rate is the DPM. The DPM
is in turnaffected by the flow type and the flow speed[8-11]. Therefore, the fan speed is
always set to high. Another factor thateffect the transmission rate is the sensor response and
resume timesdiscussed in the previous section. Finally, many other factors suchas the
environmental noise (e.g. random flow patterns in the room)can also effect the transmission
rate.Various transmission rates were tried; from one bit per 5 seconds(a character per 25
seconds) to one bit per 2 seconds (a characterper 10 seconds). To measure the reliability at
each rate, multiple experiments were performed at different separation distancesbetween the
transmitter and the receiver. Eachtransmission rate was classified at each separation distance
accordingto thefollowing ranking: very reliable (bit error rates of less than 0.01), reliable (bit


error rates of 0.01 to 0.03), unreliable (bit error ratesgreater than 0.03). Table1 summarizes
the results.

Table 1.Different transmission rates and their reliability.

At one bit per 2 seconds the transmission is unreliable at smalldistances of up to 2
meters, because of the sensors resume time athigher concentrations is longer. Moreover, at
larger distances(greater than 2 meters) successful communication is not possible atthe rate of
one bit per 2 seconds. At the rate of one bit per 5seconds, the transmission is very reliable
over various separationdistances from 4 meters to 1 meters. Based on experiments,
thesmallest transmission rate that is reliable at distances up to 4meters is one bit per 3
seconds. At this rate over the separationdistance of 4 meters the communication session is
reliable, over theseparation distance of 3 meters the communication session is veryreliable,
and over the separation distance of 2 meters thecommunication session is reliable. The reason
that the communication channel degrades slightly as the separation distance isdecreased from
3 meters to 2meters is because ofthe higher concentration levels at the sensor, and hence
longer sensor resumetimes.

This transmission rate (one bit per 3 seconds) was chosen. Further the demodulation
anddetection algorithm for this rate are discussed in detail.

3.3 Detection and Demodulation Algorithm:

Although this algorithm is slightlydifferent for eachtransmission rate, the same
underlyingprincipal is used for detection and demodulation at all rates:

1.The rate of change in the concentration at the sensor

To fine tunethis algorithm a 26 bit test sequence 10101100111000101011110110
is transmitted at the distance of 4 meters away,and the sensor reading is recorded. Figure 7
plots the sensorvoltage reading during this transmission session. Dashed red linesare used to
represent the start and the end of each bit. From thisplot a simple detection and demodulation
scheme was devised.

Thedifference between the voltage level (there are 1024 levels in theArduinos 10 bit analog
to digital converter) at the end of a bitinterval and the middle of a bit interval is measured. If
thedifference is greater than 2.2 levels (this threshold is derivedthrough experimentation), the
bit is detected as 1; otherwise the bitis detected as 0.

Using this scheme, the testphrase O CANADA (the name of the national anthem
ofCanada) was successfully sent from the transmitter to the receiver.Figure 8 shows thistest
phrase at the transmitter and received at the receiver.


Figure 7. Received signal when a 26 bit test sequence is transmitted

Figure 8. Transmitted and received text message.


4.Conclusions and Future Work:

The first known platform capable oftransmitting short text messages using chemical
signals was developed. The goal to keep the communication protocol and algorithms
simplesuch that other researchers from a wide variety of different fields ofscience could
replicate these results was attained. Moreover, the sensors andequipment were chosen to be
inexpensive and widely available.Another purpose of this platform was to motivate future
research and bridge the gap between theoryand practice.After carefully selecting the
necessary materials for this platform, the overall system response of our setup was analysed.
The linear relationship between flow speedand the delay to system response peaks maximum
was shown. The linear relationship between the flow speed andthe peaks maximum to fullwidth at half max was also shown.

Another major finding was the nonlinearity of this platform.This finding is very
important because most of current communicationtheory is based on linear systems. This
motivates further study on the exact cause of thenonlinearity in future works. If it is shown
that the nonlinearity ispart of practical molecular communication systems (i.e. the
nonlinearity cannot be resolved using better equipment), new communication theoretic work
may be necessary on this topic.


[1] NarimanFarsad et al. , Tabletop Molecular Communication: Text Messages through
Chemical Signals, 2013.

[2]Akyildiz IF, Brunetti F, Blazquez C (2008) Nanonetworks: A new communication

paradigm. Computer Networks 52: 22602279.

[3]Alberts B, Johnson A, Lewis J, Raff M, Roberts K, et al. (2007) Molecular Biology of the
Cell. Garland Science, 5 edition.

[4]Agosta WC (1992) Chemical Communication: The Language of Pheromones. W H

Freeman & Co, first edition edition.

[5]Hiyama S, Moritani Y, Suda T, Egashira R, Enomoto A, et al. (2005) Molecular

communication. In: Proc. of 2005 NSTI Nanotechnology Conference. pp. 391 394.

[6]Eckford AW (2007) Nanoscale communication with brownian motion. In: Proc. of 41st
Annual Conference on Information Sciences and Systems. Baltimore, MD, pp. 160165.

[7]Hiyama S, Moritani Y (2010) Molecular communication: harnessing biochemical

materials to engineer biomimetic communication systems. Nano Communication Networks 1:

[8]Nakano T, Moore M,Wei F, Vasilakos A, Shuai J (2012)Molecular communication and

networking: Opportunities and challenges. IEEE Transactions on NanoBioscience 11: 135

[9]Farsad N, Eckford A, Hiyama S, Moritani Y (2012) On-Chip Molecular Communication:

Analysis and Design. IEEE Transactions on NanoBioscience 11: 304314.

[10]Atakan B, Akan O, Balasubramaniam S (2012) Body area nanonetworks with molecular

communications in nanomedicine. IEEE Communications Magazine 50: 2834.

[11] Purnamadjaja AH, Russell RA (2005) Pheromone communication in robot swarm:

necrophoric bee behaviour and its replication. Robotica 23: 731742.

[12] Pedro Sousa ATdA Lino Marques (2008) Volatile marks for robotics guidance. In:
Proceedings of the EURON/IARP International Workshop on Robotics for Risky
Interventions and Surveillance of the Environment. Benicassim, Spain.

[13] Sousa P, Marques L, De Almeida A (2008) Toward chemical-trail following robots. In:
Seventh International Conference on Machine Learning and Applications (ICMLA08). pp.
489494. doi: 10.1109/ICMLA.2008.133.

[14] Munoz L, Dimov N, Carot-Sans G, Bula WP, Guerrero A, et al. (2012) Mimicking
insect communication: Release and detection of pheromone, biosynthesized by an alcohol
acetyl transferase immobilized in a microreactor. PLoS ONE 7: e47751.

[15]MooreMJ, Suda T, Oiwa K (2009)Molecular communication: modeling noise effects on

information rate. IEEE Transactions on NanoBioscience 8: 169180.

[16]Eckford AW, Farsad N, Hiyama S, Moritani Y (2010) Microchannel molecular

communication with nanoscale carriers: Brownian motion versus active transport. In: Proc. of
2010 IEEE International Conference on Nanotechnology. Seoul, South Korea, pp. 854858.

[17] Nakano T, Liu JQ (2010) Design and analysis of molecular relay channels: an
information theoretic approach. IEEE Transactions on NanoBioscience 9: 213 221.

[18]Mahfuz MU, Makrakis D, Mouftah HT (2010) On the characterization of binary

concentrationencoded molecular communication in nanonetworks. Nano Communication
Networks 1: 289300.

[19]Pierobon M, Akyildiz IF (2010) A physical end-to-end model for molecular

communication in nanonetworks. IEEE Journal on Selected Areas in Communications 28:

[20]Farsad N, Eckford AW, Hiyama S, Moritani Y (2011) Quick system design of vesiclebased active transport molecular communication by using a simple transport model. Nano
Communication Networks 2: 175188.

[21]Guney A, Atakan B, Akan OB (2011) Mobile Ad Hoc Nanonetworks with CollisionBasedMolecular Communication. IEEE Transactions on Mobile Computing 11: 353366.

[22]Srinivas KV, Eckford A, Adve R (2012) Molecular communication in fluid media: The
additive inverse gaussian noise channel. IEEE Transactions on Information Theory 58: 4678

[23]Farsad N, Eckford A, Hiyama S (2012) Modelling and design of polygon-shaped kinesin

substrates for molecular communication. In: Proc. of 12th IEEE Conference on








[24] Kim NR, Chae CB (2012) Novel modulation techniques using isomers as messenger
molecules for molecular communication via diffusion. In: Proc. of 2012 IEEE International









[25]Pierobon M, Akyildiz I (2013) Capacity of a diffusion-based molecular communication

system with channel memory and molecular noise. IEEE Transactions on Information Theory
59: 942954.
[26] OGrady M (2012) Sms usage remains strong in the us: 6 billion sms messages are sent
each day. Forrester Blogs.

[27] The International Telegraph and Telephone Consultative Comittee (CCITT) (1988)
International Telegraph Alphabet No. 2. Blue Book, Fascicle 71. 28. Bochenkov V, Sergee
VG (2010) Sensitivity, selectivity, and stabilityof gassensitive metal-oxide nanostructures.
Metal Oxide Nanostructures and their applications.