Measuring indoor air pollution: Why low-cost technology is critical
The second wave of the Covid-19 pandemic seems to be weakening. However, an old challenge remains: The heavy load of particulate matter (PM) pollution in cities. This is not good news in the Covid-19 era. Several peer-reviewed scientific studies have established that the coronavirus may stay suspended in the atmosphere as PM provides a convenient surface for the pathogen to latch on.
Therefore, managing India’s PM 2.5 crisis (indoors and outdoors) is an essential step. Unfortunately, indoor pollution and its impact of people’s health is often overlooked. But research shows that indoor air pollution, which is generated from kitchen fumes and household dust, can be two to three times deadlier than outdoor pollution.
The challenge is establishing a dense and country-wide network of indoor air pollution monitors, which meets the requirements of low cost, accuracy, and consistency of data. The continuous ambient air quality monitoring stations (CAAQMS) monitors used by the pollution control boards are not suitable for mass deployment as the Beta Attenuation Monitoring protocol they operate on is very expensive. Each unit costs more than ₹15 lakh, and with annual maintenance cost, it touches almost ₹20 lakh.
There are low-cost sensors (LCSs), which can be installed in households; however, their accuracy varies with the sensor’s sensitivity and the location of installation. For instance, if a PM 2.5 monitor does not have access to unrestricted airflow in the kitchen, it will not give accurate reading.
However, Indian Institute of Technology-Kanpur has collaborated with the Maharashtra Pollution Control Board (MPCB) to develop a new technology that provides acceptable levels of accuracy at a much lower cost.
The project, operationalised in the Mumbai Metropolitan Region, has installed 40 LCSs at 15 locations (alongside regulatory-grade monitors). Four indigenous start-ups have developed the sensors: Respirer Living Sciences, Airveda Technologies, Personal Air Quality Systems (PAQS), and Oizom Instruments, and they cost around ₹60,000 per unit. The project recorded PM 2.5 data from November 2020 to May 31, 2021.
The research team will deploy machine learning to measure and minimise the mismatch in the sensors’ readings with the co-located CAAQMS monitors. An already encouraging early finding is that their monthly uptime was 90%-plus. This is good or even better than CAAQMS units, and these monitors need minimal manual operation. Once the calibration model is complete, LCSs will be re-installed at the exact locations to improve their precision, aided by machine learning.
These low-cost sensors can be quickly expand due to a unique feature called “transfer learning”, under which a calibration model perfected in one region can be encoded into a new set of sensors across the country without repeating the process.
The project thus brings together all the desirable attributes needed to inform the suitable policy measures. In the next phase, the research team will upgrade the sensors in other ways that will help the medical community.
SN Tripathi is steering committee member, National Clean Air Programme, MoEFCC, and Vipul Arora is an assistant professor, IIT Kanpur
The views expressed are personal
Please sign in to continue reading
- Get access to exclusive articles, newsletters, alerts and recommendations
- Read, share and save articles of enduring value