SmartAir: AIoT-Based Indoor Air Quality and Occupancy-Aware Ventilation System for Good Health and Well-Being

Abstract

Affordable air quality monitoring and air quality management systems are becoming increasingly crucial for preserving public health and improving indoor environmental conditions, as well as promoting energy efficiency and climate action in buildings. An AIoT-based indoor air quality and occupancy-aware ventilation system, SmartAir, whose goal is to be a cost-effective and affordable embedded solution for demand-based ventilation, intelligent air quality management, and continuous air quality monitoring. SmartAir integrates low-cost sensors with lightweight on-device machine learning on an ESP32-class microcontroller. Two compact decision-tree models are trained and exported for embedded inference: a three-class indoor air quality classifier derived from pollutant-threshold labelling rules, and an occupancy classifier trained from environmental features to enable occupancy-aware control. The system also publishes sensor readings, model outputs, and control status to an MQTT-based data pipeline for time-series storage and dashboard visualisation. Experimental evaluation on public datasets demonstrates that the chosen models are accurate yet computationally feasible for real-time embedded execution, enabling scalable indoor deployments aligned with Sustainable Cities and Communities and healthier indoor environments.