At the core of IoT-based air quality monitoring systems are sensors that detect various pollutants. These include particulate matter (PM2.5, PM10), gases such as NO2, O3, Sulfur Dioxide (SO2), Carbon Monoxide (CO), and Volatile Organic Compounds (VOCs). Each sensor is designed to measure specific types of pollutants and is often deployed in a distributed network to cover various parts of a city or urban area. The sensors are typically compact, low-cost, and energy-efficient, making it feasible to deploy them in large numbers across urban environments, including on rooftops, street poles, and even mobile platforms like vehicles [6].

The data collected by these sensors includes real-time measurements of pollution concentrations, which are then transmitted to a central database via wireless communication networks such as Wi-Fi, 4G/5G, Long Range Wide Area Network (LoRaWAN), or Narrowband IoT (NB-IoT). The key advantage of IoT sensors is their ability to provide continuous, real-time data from multiple locations, enabling a more detailed and responsive air quality monitoring system [7].

A reliable network infrastructure is crucial for IoT-based air quality monitoring systems to function effectively. The choice of communication technology depends on factors such as coverage area, data transmission rate, and power consumption.

Wi-Fi: suitable for short-range communications, often used in dense urban areas with high internet connectivity.

Long-Range Wide-Area Network: used for long-range, low-power applications. It is particularly useful for smart city applications where sensors are deployed over large areas.

NB-IoT: a cellular network technology specifically designed for IoT devices, offering wide coverage and low power consumption, ideal for deploying sensors in remote or difficult-to-reach urban areas.

These networks enable the real-time transmission of sensor data to cloud platforms, where the data is stored, analyzed, and used to generate insights into air quality trends and pollution hotspots [8].

The volume of data generated by IoT sensors requires robust data processing capabilities. Cloud platforms that aggregate data often use AI algorithms to analyze air quality data in real time. This argument enables the detection of pollution patterns, the identification of pollution sources, and the prediction of air quality trends. Machine learning models are trained using historical data to improve the accuracy of pollution forecasts. In addition, real-time data analytics can trigger immediate responses, such as public health alerts or traffic management interventions, when pollution levels exceed safe thresholds [5].

Edge computing, a technique where data is processed closer to the sensor or at the "Edge" of the network, is also becoming increasingly important. Edge computing reduces latency by minimizing the time it takes to transmit data to centralized cloud platforms, enabling faster decision-making and reducing network congestion. It is especially useful in urban environments where quick responses to pollution spikes are necessary [9].

One of the greatest strengths of IoT in air quality monitoring is its scalability. IoT systems can be easily expanded to cover wider areas, add new sensors, or integrate additional data sources (e.g., weather or traffic data) to enhance the system's accuracy and insights. This flexibility allows cities to tailor air quality monitoring solutions to their specific needs, ranging from small-scale community deployments to citywide networks [10]. Moreover, IoT-based air quality monitoring is cost-effective compared to traditional monitoring stations, which are expensive to install and maintain. With IoT, the cost per sensor is relatively low, enabling the deployment of numerous sensors across different parts of a city to create a dense monitoring network that provides comprehensive air quality data [11].

Machine learning, a subset of AI, is instrumental in developing predictive models for air quality. These models analyze historical air quality data alongside related environmental factors, such as meteorological conditions (temperature, humidity, wind speed), traffic patterns, and seasonal variations, to learn the complex relationships that influence pollution levels. Regression models, time-series analysis, and neural networks are commonly used to predict short-term and long-term pollution trends. For instance, time-series forecasting methods, such as Autoregressive Integrated Moving Average (ARIMA) models, are often used to predict daily or hourly fluctuations in pollutant concentrations. On the other hand, deep learning models, such as Recurrent Neural Networks (RNNs) and Long Short-Term Memory (LSTM) networks, can capture nonlinear temporal dependencies in large datasets, leading to more accurate and reliable forecasts. AI-powered prediction models enable city planners and policymakers to forecast high-pollution days, anticipate potential health hazards, and take preventive measures to mitigate their impact on public health. Such predictions can be integrated into early warning systems that alert residents and vulnerable populations, such as older people and those with respiratory conditions, to avoid outdoor activities during high-pollution events [13].

Air quality data is collected continuously through IoT networks of sensors placed across urban areas. The sheer volume of data produced requires real-time processing capabilities, and AI algorithms excel at analyzing data on the fly to detect patterns, anomalies, and emerging trends. Anomaly detection models are particularly useful in air quality monitoring, as they can identify sudden spikes in pollution levels that may indicate hazardous events, such as industrial accidents, fires, or traffic congestion. These real-time insights enable authorities to respond promptly to pollution events, reducing the potential harm to public health. Real-time data analysis also empowers dynamic decision-making in cities. For instance, if AI models detect rising levels of NO2 or particulate matter (PM2.5) in a specific area, traffic flow can be temporarily altered or public alerts issued to mitigate exposure. By automating such decisions, AI helps cities manage air quality more effectively while minimizing human intervention and potential delays [14].

AI plays a critical role in identifying and tracking pollution sources. Using classification and clustering techniques, AI can analyze spatial data to identify the sources of specific pollutants, distinguishing between traffic-related pollution, industrial emissions, and natural sources such as dust or pollen. Machine learning algorithms, such as Support Vector Machines (SVM) or decision trees, can classify data and identify correlations between specific activities (e.g., high traffic congestion or industrial operations) and pollution levels, offering insights that help policymakers. Target the root causes of pollution [15].

AI-driven air quality analysis benefits greatly from edge computing, a technique that allows data processing to occur at the “Edge” of the network, closer to the source (i.e., the IoT sensors), rather than relying solely on cloud computing. This approach reduces latency, speeds up data processing, and lowers network infrastructure strain, making air quality monitoring systems more responsive and scalable. Edge devices equipped with AI can process and analyze data locally, triggering immediate alerts for high pollution levels or abnormal events. It is particularly useful in dense urban settings, where prompt responses to pollution spikes are essential [9].

AI-based air quality predictions and insights significantly affect public health and urban management policies. By providing accurate and timely data, AI helps governments enforce air quality standards, develop evidence-based policies, and allocate resources efficiently to reduce pollution. Additionally, AI-driven tools, such as mobile applications that inform residents of current and predicted air quality, enhance public engagement, enabling people to make safer lifestyle choices [12].

IoT sensors strategically deployed throughout urban environments to continuously monitor air quality are at the core of AI-IoT integration. These sensors can measure various pollutants, including particulate matter (PM2.5 and PM10), NO2, SO2, O3, and CO. Their affordability and energy efficiency make it feasible to install them in large quantities across diverse locations, such as residential areas, industrial zones, and transportation hubs [6].

The real-time data collected by these sensors is transmitted wirelessly via various communication networks— such as Wi-Fi, LoRaWAN, or cellular networks—to cloud-based platforms, where it is aggregated for further analysis. This distributed network of sensors allows for a more comprehensive understanding of air quality dynamics across the urban landscape [7].

Once the data is collected, AI algorithms play a crucial role in transforming this raw information into actionable insights. AI can use machine learning techniques to analyze historical data with real-time sensor inputs to identify trends, patterns, and anomalies. For example, regression models and time-series analysis can be employed to forecast air quality levels, helping city planners anticipate pollution spikes based on variables such as weather conditions, traffic patterns, and industrial activities. Moreover, AI can be utilized for anomaly detection, identifying unexpected pollution events that could indicate accidents, industrial leaks, or other emergencies. By recognizing these anomalies in real-time, city officials can respond swiftly, minimizing public health risks [16].

One of the most significant advantages of AI-IoT integration is the ability to perform predictive analytics. AI systems can predict future air quality trends through machine learning based on historical data and current measurements. This predictive capability empowers city officials to implement preventive measures, such as issuing public health warnings or adjusting traffic regulations on anticipated high-pollution days. By integrating AI-driven forecasts into urban planning initiatives, cities can develop more effective policies to manage air quality, including promoting green spaces, implementing low-emission zones, and optimizing traffic. This proactive approach protects public health and fosters a more sustainable urban environment [13].

Incorporating edge computing into AI-IoT systems further enhances their effectiveness. Edge computing processes data closer to the source, allowing quicker analysis and response times. This capability is especially important in urban settings, where rapid fluctuations in air quality can occur due to traffic, weather changes, or other localized events. By processing data locally, edge devices can trigger immediate alerts for high pollution levels or anomalous readings, enabling city authorities to take prompt action [17].

Integrating AI-IoT solutions also enhances community engagement by making air quality data accessible to the public. Mobile applications and online platforms can provide real-time air quality information, empowering residents to make informed decisions regarding outdoor activities, especially during pollution events. This transparency fosters public awareness and encourages community participation in air quality initiatives. Moreover, insights from AI-IoT systems can guide public health strategies, enabling authorities to identify vulnerable populations and implement targeted interventions, such as health outreach programs or pollution mitigation measures [18].

Singapore’s "Smart Nation" initiative exemplifies the successful integration of AI-IoT solutions for air quality management. The city-state employs over 100 IoT sensors distributed throughout its urban landscape to measure various air pollutants, including PM2.5, O3, and CO [22].

Data from these sensors is sent to a central platform, where AI algorithms analyze the information to forecast air quality levels and predict pollution events. For instance, during periods of haze from neighboring countries, predictive analytics enable authorities to issue timely advisories to the public, advising against outdoor activities. Furthermore, this data-driven approach supports urban planning efforts by identifying pollution hotspots, enabling targeted interventions such as establishing green spaces and improving public transport [23].

Los Angeles has embraced AI-IoT integration to address its chronic air pollution issues. The city has developed an extensive network of air quality sensors that provide localized data on street-level pollutants, enabling a detailed understanding of air quality variations across neighborhoods.

AI-driven analysis of this data has informed city policies to reduce vehicle emissions and improve air quality. For example, machine learning algorithms identify trends that help optimize traffic flow, reducing congestion in high-pollution areas. The city's air quality management system also features community engagement tools, providing residents access to real-time information and educating them on how their activities can influence local air quality [24], [25].

In response to severe air pollution challenges, Beijing has implemented AI-IoT solutions for comprehensive air quality monitoring. The city has established a robust network of sensors that continuously collect data on various pollutants, providing real-time information for analysis algorithms to forecast pollution levels and identify emission sources, facilitating informed decision-making. During periods of heavy smog, the system can issue alerts to the public, advising them to limit outdoor activities. The insights gained from AI analyses have also led to effective policy measures, such as tighter regulations on industrial emissions and incentives for adopting electric vehicles [26], [27].

Pune’s approach to air quality monitoring combines community involvement with AI-IoT technologies. The city has collaborated with local universities and NGOs to deploy a network of low-cost sensors across neighborhoods, empowering residents to actively participate in monitoring efforts. The data collected is analyzed using AI to identify pollution trends and inform local policy decisions. This community-driven approach raises awareness about air quality issues. It encourages residents to engage in initiatives to improve air quality, such as tree-planting campaigns and the promotion of cleaner transportation options [28].