1. Report on
“Mapping of carbon monoxide and green-house gases using
Wireless Sensor Networks”
Done for the IEEE-NITK Chapter Diode Group Project for the Academic Year 2013-14
(Project has been nominated for the IEEE Student Enterprise Award 2013-14)
Abstract:
A substantial section of the last decade has seen a marked responsiveness to climatic awareness. We
propose a Wireless Sensor Network as a re-configurable gas concentration monitoring solution. The
network built around the dedicated sensors will adaptively log the time scaled data. The data is
represented on pattern-observable plots. Raw data is also made available to an application developer.
Any software implementation is only a reach away from a terminal connected to the network. This
proposal gives a layered approach to the problem extending on the scalability, to support other
heterogeneous sensors and actuators, with little hardware modifications.
Statement of the Problem:
Criticality is underlined when mapping a green-house gas is being discussed. Mapping any gas
concentration is inherently a time tedious task. In a wide geographical area the concentration
variation is unpredictable; setting up locations to begin mapping, is in itself a strategic effort. In India,
no database exists for these concentration values. The ppm concentration at many urban locations
and industrial premises is alarming at many instances but is seldom known. This can be attributed to
the large-scale and human involved sampling and recording techniques majorly employed. A wireless
network as a solution to this laborious task seems inevitable.
Statement of Proposed Solution
The sensors modules designed feature 2 gas transducers which detect the concentration levels of
Carbon Monoxide. These sensors give a current output which is proportional to the concentration of
the gases. The current output from the sensors is converted into voltage and then is conditioned to
give to a value irrespective of the conditions it is deployed in. The conditioned voltage is then given to
a MSP430 for it to be converted into digital logic based on quantisation values.
The MSP430 microcontroller in each module is the interface between the sensor circuit and the
wireless module. The sensor values received are converted into digital and then sent to the XBee
module to be sent to the central node. We propose to hybridize the Zigbee and Wi-Fi approaches (IEEE
802.15.4 Zigbee Protocol and IEEE 802.11 a/b/g/n Wi-Fi/WLAN Protocols). We proceeded with RF at
Ground Zero inter-sensor-actuator communication using CC2500 module due to budget constraints.
A Raspberry Pi serves us to receive logged information. The Pi is the gateway to the server. The data
logged will be uploaded onto a server using WLAN cards/Wi-Fi or by using the Ethernet interface to
access an existing LAN.

2. Proposed Sensor Layout
A contingency takes shape. In the event of the failure of the sensor node the zonal node will
reconfigure the network using simple AT commands as UART data being passed on to the Sensor-
Xbee. This feature will thus incorporate On-the-go-programming (OTG) support. Briefly, the
communication network will function as
1. RF Front End - A star topology between the sensors and the zonal gateway Pi to begin with. The
network can always be reconfigured and hence more sensors added to the same Xbee network.
2. Zonal Node Layer - A Pi, possibly implementing an RTOS, heads the network. The zonal node can
reconfigure the network, get logged data and put it on to a server space with access through
WLAN/Wi-Fi or Ethernet.

3. 3. Server/Agent Layer - Data held in the server space will securely be handed to an authorized user. A
Publish-Subscribe model was be employed. The sensors “publish” their data to the domain space and
a client “subscribes” to it.
Research Done
Most of the research was to look into the architecture of the proposed wireless sensor network. It was
then finalised that we would proceed with the poling technique in the star-connected topology
wherein the Raspberry PI would act as the Central Hub and the MSP-430 controlled sensor nodes
would be the gateways.
The MSP430 is an ultra-low power, mixed signal micro controller from Texas Instruments. This micro
controller has the basic functionalities like Timers, Interrupts and Serial Communication which are
used in this project and thus is the processing and control unit at the sensor node end.
Research was also conducted in the protocol to be followed -
1. IEEE 802.15.4 Zigbee Protocol - Spread in the ISM 2.4GHz range of frequencies, the most commonly
used hardware employing this protocol is the XBee by Digi International. Wire/Chip/PCB/RPSMA and
three more classes, considering the transmitting power- 2mW/50mW/63mW. Data transfer rate is set
at 250Kbps.
2. IEEE 802.11 a/b/g/n Wi-Fi/WLAN Protocols - The Wi-Fi routers simply use an existing LAN
connection and convert the signals to RF and realize wireless-ness. Data rates at 54 Mbps is almost
omnipresent today.
Usefulness and Practicality
This project tends to solve the problems arising due to existing environmental degradation through
constant monitoring of harmful greenhouse effluents released using wireless sensor networks
established at appropriate points in a wide geographical area. The initial motivation was to look into
the application specifics of the project, and we stumbled upon quite a few that could prove to be
extremely beneficial to the society and humanity. The first aspect thought of is related to the
backbone of the Indian economy – agriculture. We believe that we could monitor the amount of
methane and carbon dioxide produced as by-products brought about as a result of biological
processes in plants and how the studies of their concentrations can lead to better ways to increase
the crop yield using alternative sources of energy by recycling such gases. Mapping greenhouse gases
in the atmosphere can be used to observe and predict changes in weather pattern and also to inculcate
in the minds of the younger generation, the need to innovate new ideas for conservation. The wireless
sensor network established in petrochemical refineries and chemical establishments could help in
detecting the leaks of carbon monoxide, methane and carbon dioxide based on a feedback
mechanism. The project strives to achieve characteristic differences in monitoring environments
through networks placed in an abundant of natural habitats like congested areas, canopies, coastlines,
grasslands, etc. and how differences in such environments have led to changes in pattern distribution
of wildlife.
Feasibility and Accomplishments
A lot of research has been done on the circuit design aspect by our group. A problem we faced was
the ability to condition the signal to variable parameters to produce the required concentration of the

4. particular gas for the observed voltage. We believed that the establishment of the sensor network is
a challenge that we faced. We looked to cover a wide geographical area to monitor the sensor activity
and continuous data logging on a Raspberry PI that feeds the data onto a mainframe. If given the
required funds, we think that we could look into monitoring the network from large distances to the
central hub established as part of the star-network that we developed. We looked at the Sensor
Andrew as a reference and tried to prototype our own type of protocol for secured transmission. The
scheduling and handling of events can be done by a Real Time Operating System which we tried to
work on with some help from the Computer Science Engineering students. We believe that, if given
the right amount of time and funds, each of us involved in the project could look into the application
specifics and develop our project into the domains as discussed. We are trying to imbibe in the minds
of the current generation, the need to look into developing and innovating newer ideas based on
current demographics and statistics, for a better life. We are ourselves thriving to establish the very
same and apply the developing ideas in everyday life so that the future generations could reap the
benefits of our innovation. We have currently looked into the design aspects of the individual sensors,
microcontroller coding with respect to peripheral and sensor interfacing, communicating the wireless
module with the controller and the required sequencing. The star network has been established with
the help of the PI as the central hub. We could now look into expanding our sensor network for loaded
nodes (many more sensors to monitor other parameters) so that it could prove to be useful to the
common man and the society.
Potential for Student Involvement
The potential for student involvement in this project is quite high. We already have a team of around
15 IEEE Student Members dedicated to this project who have done considerable research for its
inception and implementation. These consist of highly motivated second year and third year
undergraduate students who are willing to spare time for the implementation of this project. This
project has a very loose hierarchy, with each of the three sections having about 5 members, including
a ‘Head’ who coordinates its activities. All members have sufficient work allotted to them in order that
there is no dead weight in the group. However, implementation of this project on a larger scale will
give further scope to increase the number of students involved.
Work Completed by each team
 Analog Sensor Design Team
This team was responsible for building the analog sensor from the transducers that were
available. Due to budget constraints, we restricted ourselves to look into only carbon
monoxide sensors for now, since they were readily available and not too costly. We also
bought a ready-made sensor to check if the final expected voltage levels were the same as
that of the custom-made one that would be used as part of the sensor node. Many more gas
sensors can be made if the right transducers are available. We found that there are methane,
carbon dioxide, nitrogen dioxide, propane and oxygen transducers available from Digi-Key and
PCE Instruments. The next work that this team did after the sensors were rightly functioning
was to calibrate each of them. The calibration technique depends on individual gases and how
they respond to changes in external environment parameters. We found that carbon
monoxide being an acidic gas, has an almost direct relation between its current and PPM
values. After the sensors were calibrated, this team looked into designing the PCB for the
sensor node using the Eagle software. The sensor node was to include the MSP430, the
CC2500 and the gas sensor along with appropriate routing from each of these components.

5. The layout was printed on the PCB printing sheet; the process of imprinting and etching the
components is yet to be done.
 Interfacing and Configuring Team
This team was responsible for developing an interface between the analog sensor developed
and the control unit. ADCs, DACs, timers and counters were used in their process and most of
their work involved debugging and verification. They were also involved with understanding
the architecture of the MSP430 and reciprocate with the wireless communicating group about
the number of pins available for each type of interfacing. This team was then involved in
building and configuring the Raspberry PI. The PI was then appropriately coded and scripted
in Python to perform the tasks and schedules of poling each individual sensor nodes from a
CC2500 module readily interfaced to it. Once the data was received from the sensor node,
they were dynamically graphed using the MathLIB plot library and thus the concentration
levels of Carbon Monoxide was monitored on a real time basis from each node.
 Wireless Communication Team
This team was responsible for developing the sensor network developed. More details about
how the network was developed has been described in the section titled “The Brain of the
established Wireless Network”. This group looked into selecting the appropriate wireless
communication module depending on the application specifics. Then, modules were
developed to adhere to the group’s task of poling for data in a star network. The team then
developed libraries for interaction between the MSP430 and the CC2500 through the SPI
interface and finally to retrieve the data from the MSP and load it onto the FIFO TX buffer. The
RX buffer was supposed to wait for its call from the host before the TX buffer could send its
data to the PI. The team then focussed on developing a similar architecture at the host
interface, by developing libraries for communication between the CC2500 and the Raspberry
PI. This structure was then used as part of the algorithm/ flow-process developed by the
interfacing group. The group was inherently involved in the process of verification and
debugging based on a few test cases.
The brain of the established Wireless Network
As part of the wireless communication between the sensor nodes and the PI is concerned, we used
the TI-based CC2500 2.4GHz ISM band transceiver for low power wireless applications. We selected
this module since each sensor node will only be communicating once over a certain interval of time
with the Master through the process of Poling. The module is intended for the 2400-2483.5 MHz ISM
(Industrial, Scientific and Medical) and SRD (Short Range Device) frequency band. The RF transceiver
is integrated with a highly configurable baseband modem. The modem supports various modulation
formats and has a configurable data rate up to 500 kBaud. The CC2500 provides extensive hardware
support for packet handling, data buffering, burst transmissions, clear channel assessment, link quality
indication and wake-on-radio.
The main operating parameters and the 64-byte transmit/receive FIFOs of CC2500 can be controlled
via an SPI interface for which appropriate coding for the interfacing with the MSP430 was done. Once
the data to be sent wirelessly sits on the Transmit FIFO buffer, it is transmitted once a flag is raised
high, indicating that the master is asking the sensor node to send the data.

6. Features of the CC2500 include - high sensitivity (-104 dBm at 2.4 kBaud, 1% packet error rate), low
current consumption (13.3 mA in RX, 250 kBaud, input well above sensitivity limit), programmable
output power up to +1 dBm, excellent receiver selectivity and blocking performance, programmable
data rate from 1.2 to 500 kBaud, support for OOK, 2-FSK, GFSK, and MSK standards of communication,
suitable for frequency hopping and multichannel systems due to a fast settling frequency synthesizer
with 90 us settling time, automatic Frequency Compensation (AFC) to align the frequency synthesizer
to the received centre frequency, integrated analog temperature sensor, flexible support for packet
oriented systems (On-chip support for sync word detection, address check, flexible packet length, and
automatic CRC handling), efficient SPI interface, digital RSSI output, programmable channel filter
bandwidth, programmable Carrier Sense (CS) indicator, programmable Preamble Quality Indicator
(PQI) for improved protection against false sync word detection in random noise, support for
automatic Clear Channel Assessment (CCA) before transmitting (for listen-before-talk systems),
support for per-package Link Quality Indication (LQI), optional automatic whitening and de-whitening
of data and a few low power features which include 400 nA SLEEP mode current consumption, fast
start-up time: 240 us from SLEEP to RX or TX mode (measured on EM design), wake-on-radio
functionality for automatic low-power RX polling and separate 64-byte RX and TX data FIFOs (enables
burst mode data transmission).
In general, the CC2500 has a complete on-chip frequency synthesizer and external filters or RF
switches are not needed. It is small in size (QLP 4x4 mm package, 20 pins) and is suited for systems
compliant with EN 300 328 and EN 300 440 class 2 (Europe), FCC CFR47 Part 15 (US), and ARIB STDT66
(Japan). It supports asynchronous and synchronous serial receive/transmit mode for backwards
compatibility with existing radio communication protocols.
Technologies/ Methodologies Used
 2.4GHz ISM band RF based wireless communication between the sensor nodes and the master
 Analog Sensor Design of the Carbon Monoxide Sensor using an available transducer (other gas
sensors can also be designed with the same approach with appropriate transducers, but
calibration of PPM based on voltages is necessary)
 Calibration of the sensor which tells us the concentration of gas particles based on different
set of voltages
 PCB design for each sensor node developed using Eagle
 Configuration of the Raspberry PI and handling the poling technique for receiving data from
each sensor node
 GUI for the Raspberry PI to depict the concentration of gases at different time intervals at
different stationed positions
 Control and Processing at the MSP430 (ADC, DAC, Timer, Counter) before sending the data
through the wireless modules
References:
1) Richard Milton, Anthony Steed. “Mapping Carbon Monoxide using GPS Tracked Sensors”, 2005
URL: http://eprints.ucl.ac.uk/archive/00002179/01/emaa.pdf
2) Andrey Somov, Alexander Baranov. “Energy Aware Gas Sensing Using Wireless Sensor
Networks”,2012

7. URL:
http://disi.unitn.it/~roby/pdfs/SomovBaranovSavkinIvanovCalliariPasseroneKarpovSuchkov12EWSN.
pdf
3) Anthony Rowe, M E Berges, G Bhatia. “Sensor Andrew: Large Scale Campus wide sensing and
Actuation”, 2011
URL: www.ices.cmu.edu/censcir/resources/SensorAndrew-Tech-Report.pdf
4) “AN-1798 Designing with Electro-Chemical Sensors.” [TI Application Report], 2013
URL: http://www.ti.com/lit/an/snoa514c/snoa514c.pdf
5) “MSP430 32-kHz Crystal Oscillators”. [TI Application Report],2009
URL: http://www.ti.com/lit/an/slaa322b/slaa322b.pdf
6)
URL:
http://www.nexrobotics.com/index.php?page=shop.product_details&flypage;=flypage.tpl&product;
_id=866&category;_id=12&option;=com_virtuemart&Itemid;=45
7) URL:
http://www.digi.com/support/kbase/kbaseresultdetl?id=2213
8) URL:
http://www.digi.com/products/wireless-wired-embedded-solutions/solutions-on-module/digi-
connect/digiconnectwime#overview
9) URL: www.ti.com/lsds/ti/microcontroller/16-bit_msp430/overview.page
10) URL: http://doc.43oh.com/The_Card_Reader_SDCard_BoosterPack
11) Shane B. Eisenman, Emilian Miluzzo: “BikeNet: A mobile sensing system for cyclist experience
mapping”