REAL TIME AIR QUALITY SURVEILLANCE & FORECASTING SYSTEM (RTAQSFS) IN PUNE CITY USING MACHINE LEARNING-BASED PREDICTIVE MODEL - Research report

Aug 4, 2026

Upload by Violet

Urbanisation and industrialisation intensify air pollution, threatening human health, animal life, and vegetation while driving escalating concern for reliable air-quality measurement, monitoring, and prediction. This paper develops a Real Time Air Quality Surveillance & Forecasting System (RTAQSFS) using a cascaded architecture that integrates electronics hardware with machine-learning predictive models. Sensor-based pollutant measurements (e.g., MQ135, MQ7, MQ131) support evaluation of Linear, Ridge, Lasso, Decision Tree, and Random Forest methods using RMSE and R², showing Random Forest as the top performer with 99.9% across parameters.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages8

Reading Time20 min

AccessFree to download

Introduction

Objectives and problem context

Structure of the paper

Literature Survey

Proposed RTAQSFS System

Machine learning algorithms

Results and Discussion

Conclusion

Vol. 06, No. 2 (2024) 505-512, doi: 10.24874/PES06.02.008 Proceedings on Engineering Sciences [www.pesjournal.net](www.pesjournal.net) REAL TIME AIR QUALITY SURVEILLANCE & FORECASTING SYSTEM (RTAQSFS) IN PUNE CITY USING MACHINE LEARNING-BASED PREDICTIVE MODEL Tanuja Satish Dhope 1 Ahmed Shaikh Dina Simunic Prashant Pandurang Patil Kishor S. Wagh Sharmila K. Wagh Received 25.09.2023. Received in revised form 03.11.2023. Accepted 12.12.2023. UDC – 502.3:613 .15 Keywords: Air Quality, Machine Learning, Random Forest. Ridge, Linear Regression, IOT, Sensors A B S T R A C T Increasing urbanisation and industrialisation produce major environmental challenges such as air pollution, which endangers human, animal, and vegetation life. Reliable measurement, monitoring, and prediction of Air Quality (AQ) have emerged as key global concerns. The State GovernmentMunicipal Corporation is working on policy reforms to fight the deterioration of air quality in Pune and other Indian cities. In this paper, Real Time Air Quality Surveillance & Forecasting System (RTAQSFS) has been developed, which work in the cascaded model incorporating electronics hardware as well as machine learning algorithms. The presence of air pollutants is measured using sensors like MQ135, MQ7, MQ131 etc. The performance of machine learning algorithms viz. Linear regression, Ridge regression, Lasso regression, Decision tree and Random Forest has been evaluated wrt. RMSE and R2. The experimental results show that Random Forest outperforms the other algorithm providing less RMSE andR2 as 99.9% for all the parameters. © 2024 Published by Faculty of Engineering 1. INTRODUCTION Air pollution is frequently caused by anthropogenic activities and natural phenomena like disasters and forest fires and its levels remain dangerously high in many parts of the world. Numerous health implications of air pollution. The elderly and young children are most impacted by it. The health impacts include respiratory and cardiac disorders such as asthma, pneumonia and lung cancer. Other devastating repercussions include acid rain, eutrophication, ozone layer thinning, and global warming. Nine out of ten individuals breathe air with high levels of pollution, according to recent WHO data (Annual report, 2018) alarming 7 million people per year die from ambient (outdoor) and residential air pollution, according to updated estimates It is also growing with the increase in traffic especially in big cities. In India nearly 12.4 lakhs people lost their life due to air pollution in 2017 (Annual report 2017) .The survey indicates that in globe around 18% people loses their life due to air pollution related diseases. In India this is 26%(Annual report, 2017) . As shown in Figure 1, the Air Quality (AQ) in Pune city, Maharashtra, India is also deteriorating day by day especially in Sub-Urban regions like Shivaji Nagar, Kothrud, Mandai, Hadapsar. The environment status report (Environment status report, 2017) 2016–17 of the Pune Municipal Corporation (PMC), makes it abundantly evident that air pollutants like nitrogen oxide, nitrogen dioxide, sulphur dioxide, carbon monoxide, and particle matter are constantly rising, Hadapsar and Navipeth are two of the most polluted areas in the city. The State Government-Municipal Corporation is working on policy reforms to fight the deterioration of air quality in Pune and other Indian cities. The U. S. Environment Protection Agency (EPA) established the Air Quality Indicator (AQI) to determine the regional air quality representing pollutant allowable levels that have been inflated by time, place, and other factors (Technical assistance document, 2018) .In paper (Gulia et al,202), Respirable Suspended Particulate Matter (RSPM) has been analyzed for Delhi, India from 2003 to 2019,which shows the annual average increment in the range of 0.98–3.19% and for NO2 ,it is of 5.21–6.07% . This alarming condition needs to address by developing a s

FAQs

What system does the paper propose for Pune’s air quality?

It proposes a Real Time Air Quality Surveillance & Forecasting System (RTAQSFS) that uses a cascaded model combining electronics hardware with machine-learning algorithms for monitoring and prediction.

Which sensors are used to measure air pollutants in the system?

The paper mentions MQ135, MQ7, MQ131 and similar sensors to measure the presence of air pollutants.

Which machine learning model performs best, and how is performance evaluated?

Random Forest outperforms the other evaluated models. Performance is measured using RMSE and R², where Random Forest yields lower RMSE and R² values (reported as 99.9% for all parameters).

Aug 4, 2026

Upload by Violet