AI and low-cost sensors map air pollution in South Africa ^{CEFR B1}

Scientists have adapted experimental particle physics methods to monitor air quality in real time in South Africa. The system, called AI_r, combines a network of low-cost sensors in pollution hotspots with Internet of Things technology and artificial intelligence. The developers say this removes the need for expensive air quality monitoring infrastructure.

Air pollution is a major global health risk. The World Health Organization estimates it kills almost 7 million people every year and that 89 per cent of these deaths are in low- and middle-income countries. The World Bank finds poor health from outdoor air pollution costs up to US$6 trillion every year. Bruce Mellado, lead researcher of the South African Consortium of Air Quality Monitoring and a member of the CERN team that discovered the Higgs boson, said air quality now kills more people in Africa than HIV, malaria and tuberculosis.

The first deployment of AI_r began over a year ago as a small pilot. Several sensors were placed in schools in Soweto and the pilot showed how local pollution spikes can travel across a city. Mellado noted that a single burn can affect air quality as far as ten kilometres away. The project now plans a dense, real-time map: over the next year 500 sensors will be placed across Sedibeng, an area home to around 1 million people. The team has government support and received the ODESS prize to be presented in October, but further funding is needed to develop a continental approach.