Vaayu

Inspiration

Air pollution is not only a city-level problem; people are exposed to polluted air every day in places such as roads, housing societies, educational campuses, hospitals and public spaces. We noticed that most conventional air purifiers are designed for indoor environments, while large outdoor purification systems can be expensive and difficult to deploy widely. This motivated us to develop Vaayu — a scalable outdoor air-purification concept that combines physical filtration with real-time air-quality monitoring and data-driven analysis.

What it does

Vaayu draws polluted ambient air through a multi-stage filtration system consisting of a metal mesh filter, pre-filter, activated carbon filter and H13 HEPA filter. The system is designed to reduce particulate pollutants and selected gaseous pollutants through different filtration and adsorption stages. Vaayu also integrates air-quality sensors to collect environmental data such as particulate matter and gaseous pollutant levels. The collected data can be processed and displayed through a digital platform, enabling real-time monitoring, historical analysis and AQI-related insights.

How we built it

We first developed a working prototype to validate the physical filtration concept and airflow arrangement. The prototype uses commercially manufactured filter modules including a metal mesh filter, EU4 pre-filter, activated carbon filter and H13 HEPA filter. We designed the enclosure and airflow path around these filtration stages and incorporated a fan-based air movement system.

Alongside the hardware, we planned the software architecture around sensor-data acquisition, processing, storage, visualization and AI/ML-based AQI analysis and forecasting. The planned software stack includes Python, Scikit-learn/TensorFlow, FastAPI, React.js and MySQL.

Challenges we ran into

One of the major challenges was designing an outdoor purification system that can handle polluted ambient air while maintaining sufficient airflow through multiple filtration stages. Adding more filtration stages increases resistance to airflow and therefore requires careful fan selection and system design. Another challenge was addressing gaseous pollutants such as CO, CO₂, NO and NO₂, because conventional HEPA and activated-carbon filters alone cannot efficiently remove all of these gases. We therefore treated advanced gas-treatment stages as an area requiring further experimentation and validation rather than making unverified performance claims.

Accomplishments that we're proud of

We developed a functional Vaayu prototype and established a multi-stage filtration architecture. We also moved beyond a simple air purifier concept by designing Vaayu around real-time environmental sensing and a future data platform. The project has evolved into a scalable concept where multiple units can potentially be deployed across high-exposure areas and connected through a common monitoring system.

What we learned

We learned that outdoor air purification is considerably more complex than simply adding more filters. Filter efficiency, airflow, pressure drop, fan capacity, pollutant characteristics, energy consumption and maintenance all have to be considered together. We also learned that AI/ML can add value through analysis and forecasting, but reliable sensor data and proper validation are essential before making predictions or claiming purification performance.

What's next for Vaayu

The next stage is to conduct controlled and real-world field testing of Vaayu to measure the actual reduction of particulate and gaseous pollutants. We plan to optimize the airflow and filtration architecture, investigate suitable gas-treatment media for pollutants that are not adequately handled by the existing filters, and improve the sensor and data platform. The long-term goal is to develop Vaayu as a scalable network of smart outdoor air-purification units for locations such as campuses, housing societies, hospitals, parks and pollution-prone urban areas.

Built With

• activated
• ai/ml
• aqi
• carbon
• esp32
• fastapi
• h13
• hepa
• iot
• monitoring
• mysql
• python
• react.js
• scikit-learn
• sensors
• tensorflow