Imagine a world where you wake up every day knowing that the air you’ll be breathing is pure and clean, a world where you understand exactly what you’re breathing and where you can take meaningful action against air pollution, a world where you can plan your day around choosing the best routes for your health and are empowered with this knowledge to make your life the best it can be. Here at Managair, we believe we can turn that vision into a reality.

An increasing number of people are living in cities. By the year 2030, two-thirds of the world’s population will live in cities. Tackling the numerous problems that might arise in such environments, particularly air pollution and air-quality monitoring, is of paramount importance, and will be vital for generations to come. At the same time, there have been tremendous and rapid innovations in sensor technology, smart-monitoring systems and machine learning capabilities. We intend on leveraging these advances to build Managair, a smart, data-driven, end-to-end air-quality monitoring and recommendation system that will improve the lives of billions of people all over the world.

At a very high-level, Managair consists of a smart network of connected, smart air-quality sensor packs that can learn pollution patterns and provide personalized recommendations to users regarding air pollution in their environment. Our central Managair platform will utilize sophisticated machine learning algorithms to generate pollution models of the environment and provide personalized recommendations to users of Managair through a very simple phone app. Envision a world where you wake up, open up your phone, and just like a news broadcast, you get a list of routes that tell you the best commute route to take to avoid breathing polluted air. Providing real-time, personalized air quality readings to users is our main goal at Managair.

Managair has several innovations across a variety of levels. At the hardware level, Managair will utilize the most rugged and best air-quality sensors available to deliver an entire pollution profile around an individual sensor pack. At the systems level, we will leverage our research group’s expertise in developing robust and efficient embedded solutions to ensure that the sensor pack survives well in adverse environmental conditions while maintaining a long battery life and being lightweight. A lightweight, portable sensor pack the size of a small cell-phone will allow users to carry the pack around with them or allow for deployment on a drone, thereby allowing for a rugged lightweight and portable data gathering solution.

At the network level and higher, its Wi-Fi and GPS ranging capabilities will allow us wireless communications with Managair-enabled cell-phones while still preserving user privacy. This connectivity will enable gathering of a wide range of pollution data for an entire region, such as a neighborhood or a city or a district. GPS and Wi-Fi based localization algorithms will provide spatial resolution for multiple regions. Data-logging to a central server will provide a rich database of the pollution profiles for an entire city, and by leveraging our research group’s expertise in machine learning we will develop powerful neural network and machine learning models to enable deep insights and analytics on air quality at a personalized level as well as a very large-scale, which has never been done before.

The pilot deployment of Managair will be done in 3 stages. The first stage will consist of developing Managair sensor packs in a square mile around the University of California, San Diego campus. The packs will be spread across campus in numerous environments, and the first stage will consist of gathering pollution data across campus and generating an air pollution profile that evolves in time. Privacy preserving techniques such as differential privacy and encrypted communications will ensure user-level privacy is not compromised. The second stage will consist of using the dataset to train our machine learning models and inference engines that can provide unique insights about the air quality within an environment, such as the time of day to avoid a certain region or estimating the amount of pollutants at a spot or a path at a certain time. The third and final stage will consist of connecting individual users with the machine learning models that have been developed to provider personalized air pollution quality measurements.

Managair sensor packs will be easily manufacturable and cost-effective. We will use commodity hardware to develop the individual sensor packs and utilize mass 3D printing to create custom cases that can offer resilience against adverse environmental conditions. Managair’s power will come from a USB-powered LiPo rechargeable battery. Additionally, Managair has tremendous potential for impact across the world. The only necessary requirements for its operation are proximity to a smart phone with an internet connection and a USB power supply. The most incredible part of Managair is the more devices are deployed, the better its performance becomes. The more data is gathered and more air pollution profiles are generated, the better the quality performance of prediction becomes.



Both of us have spent an extensive amount of time living abroad and felt the need for a simple but effective way to monitor air pollution and air quality. Most of the available ones were either prohibitively expensive or in the early research and prototype stages and were very limited in their capabilities. We believe that by leveraging our research group's capabilities in Embedded Hardware and Machine Learning that we could seriously tackle this problem and make an impact. We've also had the idea for a smart, personalized air quality monitoring system for quite some time, and came across the IoT Innovation Challenge. We realized that this would be a perfect opportunity for us to pursue our idea and our passion of building technology that could help others at a global scale, and this motivated us to create a design for the IoT Innovation Challenge.


Voting is closed!