Proceeding International conference on Recent Innovations is Signal Processing and Embedded Systems (RISE -2017) 27-29 October,2017

An IoT Based Low Cost Air Pollution Monitoring System

Gagan Parmar, Sagar Lakhani, Manju K. Chattopadhyay

School of Electronics
Devi Ahilya University
Indore, India
gagansp12@gmail.com

Abstract— A prototype for an Environmental Air Pollution systems. In view of the ever-increasing pollution sources with
Monitoring System for monitoring the concentrations of major toxic chemicals, these systems should have the facilities to
air pollutant gases has been developed. The system uses low cost detect and quantify the sources rapidly.
air-quality monitoring nodes comprises of low cost
semiconductor gas sensor with Wi-Fi modules. This system Using laboratory analysis, conventional air automatic
measures concentrations of gases such as CO, CO2, SO2 and NO2 monitoring system has relatively complex equipment
using semiconductor sensors. The sensors will gather the data of technology, large bulk, unstable operation and high cost. High
various environmental parameters and provide it to raspberry pi cost and large bulk make it impossible for large-scale
which act as a base station. Realization of data gathered by installation. This system can only be installed in key
sensors is displayed on Raspberry pi 3 based Webserver. A monitoring locations of some key enterprises, thus system data
MEAN stack is developed to display data over website. The is unavailable to predict overall pollution situation. To
fundamental aspect of proposed work is to provide low cost overcome defects of traditional monitoring system and
infrastructure to enable the data collection and dissemination to detection methods and to reduce test cost, this work proposes a
all stakeholders. method combining IoT technology with environment
monitoring.
Keywords— Internet of Things, Pollution Monitoring, ARM
Microcontroller, Nucleo F401RE, Raspberry pi-3, MEAN Section II gives a brief overview of IoT architecture and
Stack how each layer is realized. Section III describes the overall
system architecture and hardware used. Experimental setup and
I. INTRODUCTION results are described in section IV. The work is concluded in

O VER the past quarter century, there has been an section V.

exponential growth of industries. These industries have
caused complex and serious problems to the
environment. Considering the significance of air quality on II. RELATED WORK
human lives, the World Health Organization (WHO) has Monitoring environmental conditions in homes have been
developed guidelines for reducing the health effects of air
inspected in [4]. A framework is proposed by author to
pollution on public health by setting the limits of the
concentrations of various air pollutants, some of which are monitor temperature, humidity and light intensity, which is
ground–level ozone (O3), nitrogen dioxide (NO2), and sulfur based on a combination of pervasive distributed sensing units,
dioxide (SO2).The first and the foremost is the severe information system for data aggregation, and reasoning and
environmental pollution which has caused deterioration of context awareness. The reliability of the sensing information is
atmosphere, climate change, stratospheric ozone depletion, loss encouraging. Several monitoring system have been proposed
of biodiversity, changes in hydrological systems and the recently for environmental pollution monitoring. While, some
supplies of fresh water, land degradation and stress on systems of the monitoring systems are specific for monitoring of CO 2
of food producing, acid rain, and global warming [1] Stationary (carbon-di-oxide).
and mobile sources release various chemical pollutants, In [5] a monitoring system is developed which gives
including suspended particulate matter (SPM), carbon the concentration of Carbon-di-oxide of remote area. The
monoxide (CO), oxides of nitrogen (NO), oxides of sulfur (SO system also report temperature humidity and light intensity of
), lead aerosol, volatile organic compounds (VOC), and other the outdoor monitoring area. Similarly, an urban CO2
toxics. It is well known that some of these chemical pollutants monitoring system presented by author in [6]. The system
have increased the occurrence of diseases such as lung cancer, operates outdoor at an urban area around 100 square
pneumonia, asthma, chronic bronchitis, coronary artery kilometers.
disease, and chronic pulmonary diseases [2-3]. Hence, there is To monitor VOC pollution levels in indoor
a growing demand for the environmental pollution monitoring environments a low-power ZigBee sensor network is proposed

ISBN 978-1-5090-4760-4/17/$31.00©2017 IEEE

524
in [7]. A WSN based system is presented for indoor and In the present work we are using only two sensor nodes, one
outdoor air quality monitoring in [8]. An array of sensors is node with MQ7 and other with MQ135 (MQ7 is a carbon-
present at each node which is connected to the central mono-oxide sensor and MQ135 detects NH3, CO2 etc.). The
monitoring unit either hardwired or wirelessly. In [9], a real- Air quality sensor MQ 135 will monitor the proportion of
time indoor air quality monitoring system is proposed. The NH3, CO2, Benzene and smoke etc. However the sensor MQ
system has seven sensors monitoring seven different gases. 7 will monitor the CO emission in the environment. The
output will be the array of numbers that will indicate the
proportion of the gases in the environment in terms of Parts
III. PROPOSED IOT ARCHITECTURE per Million (PPM).We program the Nucleo (Sensor Node) to
According to IoT architecture, the system is mainly composed read the analog value of a Sensor plugged into a simple
of perception layer (sensing layer), network layer and circuit.
Presentation layer.
B. Realization of Network Layer
Perception The primary function of network layer is to transmit sensed
Field Sensor Network Layer data by connecting all the air sensors deployed in monitoring
area to a central server and transmitting the data perceived by
sensors to data center in real time. Transmission system is
Network built according to service oriented requirement.
Data Transmission Layer In proposed work we use low cost ESP8266 Serial-
to-Wi-Fi module. The ESP8266 has a full TCP/UDP stack
support. So, the data is transferred wirelessly in the form of
TCP packets from simple node to Gateway (Raspberry pi 3).
Data Storage (NoSql)
Application C. Realization of Application Layer
Layer The whole application layer system is mainly to process
and analyze air pollutant data, evaluate air quality and then
Data Visualization predict the trend air quality develops over a period of time in
(MEAN Stack) the future. From a functional point of view, the whole
application layer includes air quality evaluation and air
pollution forecast.
Fig. 1: IoT based System’s Integral Design Architecture
Application Layer
(HTTP, DHCP, DNS, SSL,)
A. Realization of Perceptual Layer Architecture
Perception layer mainly includes Field Sensor Network
which is based on front-end acquisition device. This sensor Web Application
network hardware platform includes low power embedded
microcontroller with onboard sensors [10].
In proposed work we use Nucleo F401REtx (a 32-Bit Data Integration
ARM Microcontroller) with semiconductor gas sensor to make
simple node that collect and transmit a limited amount of data
to a central controller or gateway (Raspberry Pi 3) which Data Visualization
provides connectivity to the Internet.
Fig. 3: Application layer
Sensing Unit
In our proposed system, Application layer consist a Base
Station (a Raspberry pi micro-computer) which handles the
data transmitted from nodes. To collect data, we implemented a
Micro-controller Communication unit TCP server over Raspberry pi using Node.js (Node.js is a
JavaScript Runtime). A NoSql Database Mongo DB is installed
in Raspberry pi to store the data. All this is achieved by the OS
that is being installed in Raspberry Pi micro-computer.
Power Unit
The base station also work as a Webserver for data
visualization. So, a Webpage is implemented using MEAN
(Mongo DB, Express, Angular.js, Node.js) stack. Nginx, a
Fig. 2: Architecture of Single Node Webserver is installed for more functionality. By using Port

525
Forwarding, a DNS is assigned to the IP of Raspberry pi. The calibration procedure has been carried out in two
Whenever a client open DNS the webserver will render a steps: First finding the initial position (zero) of the sensor.
webpage and the DNS can be accessible from anywhere in the Second determining the span of the sensor. To find the zero
world. The complete system is an application of “Internet of value of a sensor, it is required to use “zero air.” However,
things” because it uses Nucleo board based motes (nodes) for there is no established standard that defines “zero air” [11].
pollution monitoring in real time and using a wireless network So, both the sensors are calibrated in clean air at morning.
to process data in a distributed information system. This procedure was adopted as there was no way of
controlling the temperature and the relative humidity available
with us.
IV. ARCHITECTURE
In this section, we describe the different components of
TABLE I
pollution Monitoring system. The overall system architecture COMPARISON BETWEEN ANALYTICAL INSTRUMENTS AND GAS SENSORS
is depicted in figure 3. It can be seen that integration of [12], [13]
hardware and software platform is obtained by combining web
server, database server gateway from the sensor nodes Features Analytical Gas sensors
themselves. In contrast with the large scale monitoring system Instruments
at TEO, the system offered in this paper is designed for small Resolution Excellent Comparable
scale environment monitoring. Cost Very high Fair
Size Bulky Compact
Rigidity Fragile Rigid
Raspberry Pi
Process Control Difficult Easy
Mass Production Difficult Easy
Database
Measurement Instantaneous Continuous
Webserver
Yet, no gas sensors are present that are 100% discerning to a
Gateway Application
single gas. Hence, it is necessary to use instruments that
employ analytical techniques to identify gases. Examples of
such instruments are Fourier transform infrared (FTIR)
instruments, gas chromatographs, and mass spectrometers.
SERIAL SERIAL Table I presents several feature comparisons between gas
WIRELESS WIRELESS sensors and analytical instruments [14].
(Wi-Fi Module) (Wi-Fi Module)
In proposed Work we are using two low cost
semiconductor gas sensors one is MQ7 while the other one is
NUCLEO NUCLEO MQ135 (MQ7 is a carbon-mono-oxide sensor and MQ135
F401RE F401RE
detects NH3, CO2 etc.) shown in figure below.
LAB2

CO2 sensor CO sensor

Node 1 Node 2

Fig. 4: Overall System architecture

A. Gas Sensors
These form the front end of the laboratory IoT systems. These
are so called “Things” of the system. Their main purpose is to
collect environmental data from surrounding (sensors) or give
Fig. 5: Gas Sensors
out data to their surroundings (actuators).
The low cost semiconductor sensors are suitable to
use in an array form for low cost environment pollution
monitoring systems. Such an array could be enhanced with
additional temperature, pressure, and relative humidity sensors
to measure the pollutant concentrations together with other
physical parameters, with the advantage of better calibration
of the gas sensors.

526
B. Raspberry pi 3 with a Wi-Fi Module (ESP8266) each is interfaced with
Nucleo F401RE to form two client nodes. Fig 6 below shows
a client node (simple node). These Client nodes send sensor
data wirelessly from sensor node (Nucleo) to Gateway
(Raspberry pi 3) in the form of packets. These data packets
contain id of the node data and the time of packet created. To
save and analyze data we have installed a NoSql database
(Mongo DB) in raspberry pi. The data saved in JSON format
(JavaScript Object Notation).So, it is simple to render the web
page with JSON formatted data. For data visualization we
have designed a webpage using MEAN stack.
We have assigned a DNS to the static IP of
Fig. 6: Raspberry pi 3 Raspberry pi and forwarding the port over internet. So, that it
We have used Raspberry Pi 3
as a Base Station in our
system. Raspberry Pi is configured as credit size
microcomputer based on Raspbian Linux Operating System.It
offers less complexity and more affordable solutions for
wireless monitoring [15]. The Raspberry Pi 3 Model B
features a quad-core 64-bit ARM Cortex A53 clocked at 1.2
GHz ,1 GB SDRAM with built-in Wi-Fi and Bluetooth.
The board also provides user with ETHERNET
which is advantageous to our system. This results in remote
access management through SSH login by putting IP address
in it.

C. Nucleo F401REtx
Fig. 4 shows a Nucleo F401RE microcontroller has ARM®32-
bit Cortex®-M4 CPU with 84 MHz max CPU frequency, 512
KB Flash, and 12-bit ADC with 16 channels. It has Arduino can accessible from anywhere in the world. Fig 8 shows
compatible14 digital input/output pins (of which 6 can be used webpage designed using MEAN stack.
as PWM outputs), 6 analog inputs, a USB connection, an ICSP
header, and a reset button. It also contains everything needed
Fig. 8: Experimental setup & Connection
to support the microcontroller, we can connect it to a computer
with a USB cable to get started [16].

Fig. 7: Nucleo F401RE Board

V. EXPERIMENTAL SETUPS AND RESULT

Fig. 9: Webpage using MEAN stack
We start with the Raspberry pi 3 B model as base station.
Raspberry pi’s inbuilt Wi-Fi is configured in such a way that it
acts as a coordinator. VI. CONCLUSION AND FUTURE SCOPE
A TCP server is implemented over raspberry pi using We designed a low cost pollution monitoring system. The
Node.js in JavaScript. While gas sensors MQ7 and MQ135 semiconductor gas sensors can be successfully used to monitor

527
the target gas concentrations. The usage of the semiconductor (MIPRO), 2014 37th International Convention on, pp.1013, 1018,
26-30 May 2014
sensors adds several advantages to a system such as low cost,
[16] R. Bhadoriya, M. K. Chattopadhyay and P. W. Dandekar, "Low
quick response, low maintenance, ability to produce cost IoT for laboratory environment," 2016 Symposium on
continuous measurements, etc. Colossal Data Analysis and Networking (CDAN), Indore, 2016,
One major advantage provided by the system is small pp. 1-4.
size. The gateway node of Wireless Sensor Network, database
server and web server all are packed into a single compact
credit sized micro-computer Raspberry Pi. This makes the
system very portable. In addition, this system allows us to
integrate other hardware components with the Raspberry Pi as
a credit size microcomputer. System can be upgraded by
adding more sensing nodes. The detailed design and
measurement results demonstrate the usefulness of this
system. The prototype can be realized as a handheld mobile
device for pollution monitoring.
VII. REFERENCES

[1] WHO, “Global Environmental Change,” World Health

Organization, Geneva, Switzerland, 2005.
[2] Air Resource Management Centre, “Vehicle-related air pollutants
and public health,” Ministry of Environment and Natural
Resources, Sri Lanka, May 2003, pp. 6–11.
[3] Urban Air Quality Management Project, “Impact of fuel changes
on exhaust and evaporative emissions,” Ministry of Environment
and Natural Resources, Sri Lanka, May 2003, pp. 75–85.
[4] S. Kelly, N. Suryadevara, and S. Mukhopadhyay, “Towards the
Implementation of IoT for Environmental Condition Monitoring in
Homes,” Sensors Journal, IEEE, vol. 13, no. 10, pp. 3846–3853,
Oct. 2013.
[5] H. Yang, Y. Qin, G. Feng, and H. Ci, “Online Monitoring of
Geological CO2 Storage and Leakage Based on Wireless Sensor
Networks,” Sensors Journal, IEEE, vol. 13, no. 2, pp. 556–562,
Feb. 2013.
[6] X. Mao, X. Miao, Y. He, X.-Y. Li, and Y. Liu, “CitySee: Urban
CO2 monitoring with sensors,” in INFOCOM, 2012 Proceedings
IEEE, Mar. 2012, pp. 1611–1619.
[7] C. Peng, K. Qian, and C. Wang, “Design and Application of a
VOC Monitoring System Based on a ZigBee Wireless Sensor
Network,” Sensors Journal, IEEE, vol. 15, no. 4, pp. 2255–2268,
Apr. 2015.
[8] O. Postolache, J. Pereira, and P. Girao, “Smart sensors network for
air quality monitoring applications,” Instrumentation and
Measurement, IEEE Transactions on, vol. 58, no. 9, pp. 3253–
3262, Sept 2009.
[9] J.-Y. Kim, C.-H. Chu, and S.-M. Shin, “ISSAQ: An Integrated
Sensing Systems for Real-Time Indoor Air Quality Monitoring,”
Sensors Journal, IEEE, vol. 14, no. 12, pp. 4230–4244, Dec. 2014.
[10] J. Yang, C. Zhang, X. Li, Y. Huang, S. Fu, M.F. Acevedo.
Integration of wireless sensor networks in environmental
monitoring cyber infrastructure. Wireless Networks,
Springer/ACM, Volume 16, Issue 4,pp. 1091-1108, May 2010
[11] IST, “Gas sensor calibration,” International Sensor Technology,
CA, USA, Ch. 11, pp. 161–173.
[12] D. D. Lee and D. S. Lee, “Environmental gas sensors,” IEEE
Sensors J., vol. 1, no. 3, pp. 214–215, Oct. 2001.
[13] Capteur Sensors, “Principle of operation,” General Information on
the Capteur NGL07 Carbon Monoxide Sensors, Capteur Sensors,
U.K., 2000, pp. 1–13.
[14] N. Kularatna and B. H. Sudantha, "An Environmental Air
Pollution Monitoring System Based on the IEEE 1451 Standard for
Low Cost Requirements," in IEEE Sensors Journal, vol. 8, no. 4,
pp. 415-422, April 2008.
[15] Vujovic, V. , Maksimovic, M., “Raspberry Pi as a Wireless Sensor
node: Performances and constraints,” Information and
Communication Technology, Electronics and Microelectronics

528