Air pollution contributes to 7 million annual premature deaths, the rough equivalent of the population of Hong Kong. This alarming statistic has inspired many people to take action. By monitoring air quality in human settlements, researchers and citizen scientists have developed plans to help protect people against the worst effects of pollution.
Problems With Air Quality Sensors
Governments often install permanent air quality sensors throughout urban areas. However, these monitors can be few and far between, making it difficult to understand local pollution levels. Additionally, many sensors only provide real-time air quality reports, which doesn’t help with making plans based on predictions.
Satellites play a key role in gathering global air quality readings but don’t take frequent measurements. This leads them to miss localized events – such as heavy traffic or new construction projects – that temporarily worsen air quality. Their precision is also limited near the planet’s surface where people breathe.
Furthermore, they don’t take readings inside people’s homes. Indoor volatile organic compound (VOC) levels can be two to five times higher than outdoors. People also smoke, cook with propane and use kerosene lamps in their homes, which all cause significant air pollution.
How AI Can Help
Crowdsourcing projects are underway to use mobile air quality meters to gather more data. Citizen scientists are placing these mobile sensors in public places such as community centers, schools and buses to understand air pollution levels better. Some projects employ AI to decide where to take measurements – it helps reduce the number needed by identifying the most critical areas to examine.
The large volume of information gathered via these projects allows researchers to train machine learning algorithms. AI software looks for patterns in air quality measurements, current weather conditions, traffic and the built environment to predict upcoming pollution levels. People can then use apps, like IQAir, to check the forecast in various locations.
In Delhi, India, students created an app called Air Cognizer that estimates air pollution based on photos. They developed a data set of almost 5,000 images and trained the AI to recognize the color and transparency of the sky. Now, app users can simply snap a picture to get an idea of current air quality.
Air quality data helps schools, governments and individuals alter their behavior based on current and forecasted conditions. For example, employers might allow people to work from home on highly polluted days when taking public transit could pose a health risk. Hospitals can make staffing decisions based on upcoming pollution levels. Cities can warn people not to drive on days with severe traffic emissions.
Air quality data is also important for enacting regulatory changes. Policymakers who see how severe the pollution problem is in their cities may be more likely to set guidelines to limit emissions and encourage green energy initiatives. The Climate and Clean Air Coalition estimates that reducing methane emissions by 40%, hydrofluorocarbon by 100% and black carbon by 70% could save 2.4 million lives annually.
Although air quality predictions aren’t perfect, apps and citywide alerts can include an uncertainty analysis to give people an idea of the prediction’s accuracy, just like a weather forecast. This data would integrate well with smart cities using IoT to gather information and provide updates.
The Role of AI in Air Quality
Artificial intelligence can significantly affect how people measure, analyze and forecast air pollution levels. Individuals can use this data to make more informed decisions regarding staffing and when to go outside. It also helps with route planning, issuing citywide alerts and changing emissions policies. AI will play a key role in air quality management going forward.
The post Why We Need AI for Air Quality appeared first on Datafloq.
