Back to Aditya's profile
8views

Air Edge Ai

Aditya Biswal
  • Expert reviewed· 5 reviews
Air Edge Ai

Submitted to The Edge AI Air Detective

Arduino board used?

Yes

Edge AI used?

Yes

AI/ML model used

None

Sensors used

Yes

Camera used

None

Modulino used

None

Problem being solved

Existing air-quality monitoring often relies on cloud services and city-wide AQI data, which may not reflect local conditions and can be affected by network delays or outages. AirEdge AI aims to enable affordable, hyperlocal air-quality monitoring using Arduino UNO Q, local sensor processing, pollution-source identification, forecasting, and immediate actionable guidance.

AI model training platform

Not yet selected. The model training platform will be specified after the AI/ML model is chosen and implemented.

Arduino UNO Q features used

Arduino UNO Q 4GB for sensor acquisition and edge computing, with a planned Python-based data pipeline, local processing, and web dashboard integration. On-device inference and offline operation remain to be validated.

Deployment method

Prototype deployment using Arduino UNO Q 4GB, connected environmental sensors, an LCD, an LED alert indicator, and a planned local Python/Flask backend with a browser-based dashboard.

Key innovation

AirEdge AI aims to combine low-cost environmental sensors with edge computing to provide hyperlocal monitoring, reduce dependence on cloud connectivity, and support future on-device pollution forecasting, source identification, and immediate guidance.

Solution summary

AirEdge AI is an Arduino UNO Q-based environmental monitoring prototype using an MQ-135 and DHT11 to collect gas-sensing, temperature, and humidity readings. It provides threshold-based LED alerts and LCD output, with a web dashboard planned for monitoring and history. Local AI forecasting, pollution-source identification, and offline inference are future development goals.

Report (PDF)
Report (PDF)PDF
Arduino board used?

Yes

Edge AI used?

Yes

AI/ML model used

AirEdge AI is an intelligent monitoring and assistance system that combines machine learning, computer vision, and sensor-based monitoring. For hand gesture recognition, the project uses the MediaPipe Hand Landmarker, a pre-trained machine-learning model that detects 21 key points on a person's hand through a laptop webcam. The OpenCV library captures and processes the video frames, while custom Python logic interprets the detected landmarks to identify gestures. 1. ML model: MediaPipe Hand Landmarker 2. Model file: hand_landmarker.task 3. Framework: MediaPipe Tasks 4. Computer vision library: OpenCV 5. Gesture classification: Custom Python rules 6. Training: An under development training model

Sensors used

MQ-135 Gas Sensor – Air-quality and gas concentration monitoring. DHT11 Sensor – Temperature and humidity measurement. Laptop Webcam- Fall detection

Camera used

Yes used. And the camera module is a webcam.

Modulino used

None

Problem being solved

AirEdge AI addresses two key problems: indoor air-quality monitoring and assistance for people who may be unable to call for help physically. It uses an MQ-135 sensor to monitor air-quality changes and a DHT11 sensor to measure temperature and humidity. Using Arduino UNO Q, the system provides visual and audible alerts when sensor readings cross configured thresholds. Additionally, a laptop webcam with the MediaPipe Hand Landmarker model and OpenCV recognizes hand gestures, allowing a person to request help by holding up an open palm. This provides a gesture-based assistance mechanism for people who have difficulty standing or reaching a phone.

AI model training platform

Google MediaPipe

Arduino UNO Q features used

1. Dual-processor architecture – Supports microcontroller-based hardware control and Linux-based application processing. 2. GPIO control – Interfaces with the MQ-135, DHT11, LEDs, buzzer, and monitoring switch. 3. Python application support – Runs the sensor-data handling and backend communication logic. 4. Arduino sketch execution – Reads sensor values and controls hardware alerts. 5. Wi-Fi/network connectivity – Sends sensor readings to the Flask backend over the network. 6. Arduino App Lab – Used to develop and run the project application, including startup configuration.

Deployment method

AirEdge AI uses a hybrid deployment method. The Arduino UNO Q runs the hardware-control and sensor-monitoring application, collecting air-quality, temperature, and humidity readings. Sensor data is transmitted over Wi-Fi to a Flask backend running on a laptop. The backend stores readings in PostgreSQL and provides data to the web dashboard. Hand gesture recognition is deployed locally on the laptop using OpenCV and the pre-trained MediaPipe Hand Landmarker model. The system uses local processing and network communication to provide real-time monitoring and gesture-based help requests.

Key innovation

AirEdge AI combines real-time air-quality monitoring with gesture-based assistance in a single system. It uses Arduino UNO Q to monitor air-quality, temperature, and humidity, providing visual and audible alerts when readings cross configured thresholds. Its key innovation is integrating a webcam-based hand gesture recognition system using the pre-trained MediaPipe Hand Landmarker model and OpenCV, enabling users to request help by holding up an open palm without needing to stand or reach for a phone. The system combines edge-device sensor monitoring, local computer vision, and a Flask-based web dashboard for accessible, responsive monitoring and assistance.

Solution summary

AirEdge AI is an intelligent monitoring and assistance system built using Arduino UNO Q, sensors, machine learning, and computer vision. It monitors air quality using the MQ-135 sensor and measures temperature and humidity using the DHT11 sensor. The system triggers visual and audible alerts when readings exceed predefined thresholds. For assistance, it uses a laptop webcam, OpenCV, and the pre-trained MediaPipe Hand Landmarker model to recognize hand gestures. Holding an open palm generates a help request, while a fist resets the gesture state. A Flask backend and PostgreSQL database support sensor-data management and web-based monitoring. The solution combines environmental monitoring with accessible, gesture-based assistance.

Report (PDF)
Report (PDF)PDF
Schematic
Schematic: Air Edge Ai