CHENNAI, Indian Institute of Technology Madras (IIT Madras) Researchers
have developed a low-cost mobile air pollution monitoring framework in which, pollution sensors
mounted on public vehicles can dynamically monitor the air quality of an extended area at high
spatial and temporal resolution.
Traditionally, ambient air quality is measured in monitoring stations and reported as ‘Air Quality
Index’ (AQI). Since these stations are at fixed locations, they only measure the air quality of a
small geographic area.
Air pollution however is dynamic with locations just a few hundred meters away from each other
exhibiting different levels of pollution. Levels can also vary at different times of the day.
However, setting up more stations is not practical because of the high costs.
Towards tackling this issue, IIT Madras Researchers, have developed a new IoT-based mobile
air pollution monitoring technology wherein low-cost air quality sensors are mounted on vehicles
to gather spatio-temporal air quality data. For the cost of a single reference monitoring station, it
would be possible to map an entire city at high resolution using these low-cost mobile
monitoring devices.
Led by Prof. Raghunathan Rengaswamy, Dean (Global Engagement) and Faculty, Department
of Chemical Engineering, IIT Madras, Project Kaatru (air in Tamil) leverages IoT, big data and
data science to achieve the following goals:
 Obtain pan-India hyperlocal air quality map
 Exposure assessment for each Indian citizen
 Data driven solutions for policy, intervention and mitigation strategies
A data science and IoT based mobile monitoring framework for performing high resolution
spatio-temporal assessment was recently published in the reputed, peer-reviewed journal
Building and Environment (https://doi.org/10.1016/j.buildenv.2022.109597) in a paper co-
authored by Sathish Swaminathan, Anand Guntuku, Sumeer S, Amita Gupta and Prof.
Raghunathan Rengaswamy.
Elaborating on the findings of this Research, Prof. Raghunathan Rengaswamy, Faculty,
Department of Chemical Engineering, IIT Madras, said, “Interestingly, one specific location
showed a significant spike of PM2.5 pollution between 2 am and 3 am. This was associated to
trucks carrying milk from a major milk distribution hub in this location at this time. PM2.5 spikes
were also found in school neighbourhoods during school start and end hours and in commercial
zones during peak hours.”

Prof. Raghunathan Rengaswamy added, “Mobile air quality sensors would find extensive use
in both personal and public health initiatives. Personal monitoring devices can help people know
the extent of pollution in their neighbourhood so that they can take protective measures. Traffic
can be rerouted if local pollution levels are known. Government policy changes and smart city
planning would benefit enormously from the use of mobile air quality trackers. Our affordable
IoT based mobile monitoring network, coupled with data science principles offers
unprecedented advantage in gathering hyperlocal insights into air quality. It is the only viable
option at present, capable of offering high spatio-temporal awareness that could allow for
informed mitigation and policy decisions.”
The devices are capable of measuring multiple parameters, ranging from PM1, PM2.5, PM10
and gasses such as NOx and SOx. In addition to pollutants, the devices can assess road
roughness, potholes and UV index among others. The modular design of the device allows for
sensors to be replaced on demand. Figure 1 shows the parameters that can be sensed by the
IoT mobile monitoring devices.
The patented IoT side view mirror design enables the devices to be retrofitted on any kind of
vehicle, ranging from buses to cars and even two wheelers.
The IoT devices are also equipped with GPS and GPRS systems to collect and transmit location
information. Data Science principles are used to analyse the large volume of data generated
from these IoT devices.
The researchers undertook two case studies as a part of this work.
The first study was aimed at assessing hyper local air quality assessments to evaluate
the effects of vehicular traffic, urban topography and urban functions. Measurements
were made across a 15 sq. km. area of carefully selected region in western Chennai to study
and authenticate as to how pollution concentration varied.
The pilot area was chosen carefully to include different land use such as commercial, industrial,
residential, hospital and school zones, that would have an impact on unevenly distributed
emission sources, dilution, and physicochemical transformations over short distances.
Additionally, the impact of different factors such as: vehicular density, urban, industrial &
residential functions on hyperlocal level air quality were assessed. Diurnal trends in specific
zones were identified and correlated with human activity in those zones.
The study was able to capture even subtle variation in PM2.5 concentrations at various
locations across time. The gradation in PM2.5 concentration between main roads and arterial
roads was also captured through this assessment. This case study was done jointly with the
Centre for Urbanization Buildings and Environment (CUBE), a Centre of Excellence at IIT
Madras.
The second case study was the analysis of PM2.5 levels around a specific high intensity
event – Deepavali (Diwali). The auto-rickshaw-mounted sensors were made to ply on South
Chennai roads two days before Deepavali of 2019, two days of the festival, and two days after.
Four specific areas were chosen - a commercial area, a heavily wooded residential academic
campus, an upscale residential area and an industrial area with small-scale workshops.
The published Research Paper also validates the reliability of the data collected by the IoT
devices by comparing it with a CPCB station in one of the locations of study. Data collected

from the devices followed the same profile/trend across the 6 days of study and showed a high
qualitative match with the nearby CPCB station.
A detailed analysis of the variation in PM2.5 levels before, during and after the high intensity
event was done across locations and time. Through such an assessment, it was possible to
gauge the impact of the event on an area and associate it with the type of area (land use). This
insight confutes the popular opinion that the entire region experiences the same impact during
such high intensity events.

Click Here for More Exams / Admission