Indoor Air Pollution Prediction and Prevention Using IoT and Machine Learning Techniques
Abstract
Monitoring indoor air quality (IAQ) has become increasingly important as people spend a significant amount of time indoors, whether at indoor environments (home, office, Conf rooms, Auditoriums). Using a comprehensive system that integrates IoT sensors and machine learning techniques offering an effective way to ensure healthier indoor atmosphere. This system consists of several components and steps. IOT sensor devices are deployed to measure parameters such as CO2, PM2.5, PM10, VOC, Temperature and humidity concentration levels in various indoor space which includes bedrooms, living rooms and Kitchens. These sensors continuously collect real time IAQ data from the sensors. The collected data from the sensors are then transmitted to a central microcontroller device, which acts as an aggregation point and responsible for preprocessing the data, performing initial filtering, or smoothing if necessary and package it to transmit to central storage (cloud). Reliable communication protocols such as Wi-Fi, Bluetooth are used to send the data from the microcontroller to a central server for further processing. The collected data is securely stored in scalable storage solutions like cloud-based servers (ThingSpeak , AWS, Azure) or local databases ensuring the data integrity, availability and security
For User access, user friendly Grafana dashboard is developed to visualize IAQ data in real time. Authorized users can access this dashboard to monitor IAQ statistics, view historical trends and receive alerts if any parameters exceed safe readings. Machine learning algorithms are applied to analyze the IAQ trends in the data. Techniques like regression are used to predict future IQ parameters taking into consideration of historical data and different factors such as time of the day, occupancy. Different classifications algorithms categorize IAQ into levels such as good, moderate, poor along with providing appropriate recommendations.
Alerts and notifications are implemented to inform users in real time if IAQ parameters reach critical levels or if preventive actions are needed. The systems is continuously improved by collecting the user feedback, which in used to fine tune the ML models and enhance the preventive measures.
This approach can contribute to healthier and more comfortable indoor environments in homes and offices while also helping to reduce health risks associated with poor IAQ