Artificial Intelligence against air pollution in Africa thanks to the Google AI initiative
Curated by Antonio Aversano
We have known this for a long time now: the air we breathe, especially in urban centers with greater population density and in industrial areas with high environmental impact, has suffered and continues to suffer drastic drops in terms of quality, leading to an increase in diseases and victims of environmental pollution. Dr. Engineer Bainomugisha of Makerere University , together with other collaborators and students, has developed a system to monitor the level of air pollution in Kampala, Uganda, Africa, thanks to the use of Artificial Intelligence as part of the Google AI initiative .
AI against air pollution in Africa: a project funded by Google
Engineer and the Makerere University team are one of 20 organizations selected from over 2600 candidates to receive a donation through the “ Google AI Impact Challenge “ . This is a Google.org initiative to help nonprofits, startups and researchers use the power of AI to address environmental and social challenges. As part of the program, the Makerere team also received training and mentoring from AI experts from Google and DeepMind during a 9-month crash course on Artificial Intelligence.
Kampala, like many other cities in sub-Saharan Africa, has a critical lack of data on the extent and extent of air pollution. This challenge inspired Engineer and his students to create AirQo, a low-cost air quality monitoring system.
Challenges and innovation
The main factors of air pollution in Kampala are transport, industry, the combustion of wood, coal and waste. Kampala is the political capital and financial district of Uganda and contributes over 30% of Uganda's GDP. The city is home to more than 32% of the country's manufacturing facilities and therefore industrial emissions from activities, such as metalworking, furniture, textiles, and plastics, will contribute to a significant amount of air pollution.
Traditional air quality monitoring systems are very expensive . Hundreds of thousands of dollars for installation and maintenance, the latter requiring specific capabilities not available everywhere. They are also assumed to have a constant need for electricity and Wi Fi network, not to mention that they are not suitable for dusty environments.
AirQo devices, on the other hand, are specifically designed to withstand the environmental conditions of many African cities , such as dust and extreme weather conditions. They also include a wide range of power and data transmission options, so they can operate in areas where there is limited access to the power grid or poor internet connectivity.
Unlike traditional equipment that allows fixed installation in a single location, thus limiting the monitoring area, AirQo sensors are installed on mobile platforms. In fact, the Engineer and his collaborators used “mototaxis” to house the devices, in order to create driving monitoring stations, called boda boda , which widen the controlled area as the vehicles move around Kampala.
The devices continuously take air samples from one location and use a “light scattering” method to quantify the particulate concentration. More specifically, the equipment is designed to monitor PM 2.5 and PM 10 fine particles, location details, indoor and outdoor temperature, atmospheric pressure and humidity, transmitting data on the 2G network which is predominantly available in many parts of Africa.
After data collection, the team uses cloud-based artificial intelligence software to analyze particle data in the air in real time and predict local pollution levels.
Scientists hope that air quality data and information from their network can be used to inform decision-making and actions by city governments and regulators to control, better manage and improve the quality of air. air in the Kampala area and beyond.
You might be interested in: Cyborg jellyfish to monitor marine pollution
The article Artificial Intelligence against air pollution in Africa thanks to the Google AI initiative comes from TechCuE .