A IoT-Enabled Multisensor Air Quality Monitoring with Edge-Based AQI Computation and Intelligent Mobile Analytics

Inexpensive air quality monitors based on microcontroller units (MCUs) generally end at sensor data collection and numerical displays, with no further interpretation of data meaning and possible actions left to the user. The paper describes a multisensor air quality monitoring system combining an ESP32-WROOM MCU, DHT22 temperature humidity sensor, analog MQ-series gas sensor, infrared optical dust sensor and a NEO-6M GPS module together with a custom cross-platform Flutter application. Unlike cloud-based IoT systems, which depend on an active Wi-Fi/cloud connection, the sensor node communicates with the application over a dedicated Bluetooth Classic serial connection sending newline-delimited JSON payloads, which are processed by the application to display live data in a dashboard including an overall computed air quality index in addition to individual sensor values and GPS status. The application includes a feature to send an on-demand analysis of the current reading set to a cloud-based generative-AI model (Google Gemini) using a constrained prompt format and receive back a structured non-diagnostic report of environmental risks, potential causes, precaution and short-term forecast. The functional tests have confirmed the consistency of JSON framing for both USB-serial and Bluetooth-serial connections as well as correct response of the system's air quality classification algorithm from a low pollutant baseline reading to high particulates test scenario. Overall, a combination of Bluetooth serial link from a device to a phone and cloud-based generative-AI model analysis proves to be an inexpensive and feasible alternative to IoT cloud infrastructure for contextual air quality information delivery.