How we can use AI to fight air pollution

Air pollution remains a problem almost everywhere, even as other environmental issues such as global warming, biodiversity loss, soil degradation and unsustainable use of freshwater resources become more prominent, But air pollution remains a problem that deserves our attention and action.

According to the World Health Organization, 3 million to 8 million people die prematurely every year because the air they frequently breathe contains harmful substances that may affect the respiratory system, cause inflammatory diseases, or affect the human immune system .

Despite several regulations aimed at reducing the emission of air pollutants and placing limits on the concentration levels of ambient air pollutants, measurements across Europe still frequently show concentration levels exceeding thresholds for human health and food production safety.

The rest of the world has bigger problems. Sometimes, for example, in megacities in South and East Asia, Africa, and South America, pollution is so bad that people can barely work or navigate the streets.

We are therefore advised to continue and even expand the monitoring of air pollution and further develop the tools needed to analyze these measurements and make predictions about air pollutants for the benefit of vulnerable groups Get a warning and take countermeasures. In this article, we will see how artificial intelligence can be used to combat air pollution.

Artificial intelligence to prevent and control air pollution

Regarding global air pollution, we have a lot but too little data. In order to build good AI tools, AI requires large amounts of data, and it is necessary to understand what data is available and what information this data contains. Since the 1980s, air pollution monitoring networks including fixed stations and mobile platforms have been established in many regions around the world.

Although satellite instruments cover the globe, their measurement frequency is not high enough, and their measurement accuracy near the earth's surface where humans breathe air is also limited. There are almost no air quality monitoring stations in many areas of the world. Even in Europe, where the network of monitoring stations is relatively dense, there are usually ten or even one hundred kilometers between adjacent monitoring stations.

Artificial intelligence could play a role in expanding global air pollution monitoring networks, for example, as a means of interpreting measurement signals obtained from modern low-cost sensor devices. Such equipment can be used to fill monitoring gaps if used in conjunction with measurements from traditional stations.

Artificial intelligence can help analyze and predict air pollution

Explanation and prediction of air pollution currently require complex numerical models, so-called chemical transport models, which Model weather and air pollution chemistry using computer code that contains thousands of lines and runs on the world's largest supercomputers.

Using AI for these purposes presents some challenges that are different from those commonly seen in other AI applications. In the 1990s, AI methods were first tested in the context of local air quality predictions. At that time, machine learning algorithms and computing power were about a million times weaker than today, so machine learning results were only slightly better than those obtained using classical statistical methods.

After 2012, so-called convolutional neural networks achieved breakthroughs in typical artificial intelligence tasks such as image recognition, and atmospheric scientists became interested in artificial intelligence again. Since 2018, several studies have shown that advanced machine learning techniques can indeed produce high-quality air pollution forecasts locally.

Machine learning models will soon also provide alternative, computationally cheaper solutions for predicting air pollution in an area. Such systems may work best in a hybrid approach, where weather information comes from traditional numerical simulations, that is, weather forecasts, and air quality information comes from measurements.

Opportunities and Risks of Artificial Intelligence in Air Pollution Management

The combination of low-cost air pollution sensors with artificial intelligence and hybrid models, More detailed maps of air pollution may be available, and therefore more targeted mitigation measures than currently affordable measures.

Combined with physiological sensors and medical information systems, AI-based pollution monitoring may eventually enable direct measurement of doses of inhaled pollutants, helping vulnerable groups better plan their outdoor activities and avoid hazardous environments. In fact, several companies in Europe and elsewhere are already promoting AI-based air quality information.

At this point, however, the quality of such systems is often questionable, and there is little information about how well they work in practice. As in other application areas, the greatest danger with AI solutions occurs when trust is blind. Therefore, it is important that we fully understand the capabilities and limitations of AI-based air quality monitoring systems and that we always control our actions.

The above is the detailed content of How we can use AI to fight air pollution. For more information, please follow other related articles on the PHP Chinese website!