Contact

Using Artificial Intelligence to predict concentrations of air pollution

A team of researchers are working on a novel way to predict high concentrations of polluted air over cities, thus reducing exposure and hospital admissions.

Number crunching is no longer enough

In a previous article we talked briefly about how general pollution is an ugly side effect of urban and industrial growth. The production of air, soil and water contaminants is just a part of the process behind becoming a global civilization, and there doesn’t seem to be another way, at the present moment, to do it. Green technology is just barely becoming relevant enough to be noticed by the public and the big industrial players, so we should probably hold our breath if we expect to see results anytime soon. Global problems require global solutions just as much as large enough time for them to act. We won’t see any improvements tomorrow.

None of that means that we can’t do something about our present air pollution problem. Technology’s side effects may generate pollution, but the same technology can also be used to try and fix it. Or, as in the case of air pollution forecasting, inform people about concentrations of pollutants so that they can take measures to avoid harm. Air quality forecasts are good tools that can be used to understand the concentrations of pollutants in a city or region. When accurate, they help people plan ahead of time, decrease effects on health, such as asthma and upper respiratory tract infections, and reduce any costs associated to their treatment. By extension, good air quality forecast also alleviate admissions to hospitals, since an increase in air pollution lead to an increase of hospitalizations.

Traditional air pollution forecasts are created by using statistical patterns based on atmospheric chemical analysis, which are then turned into numerical models. These work as calculators, and can be used to determine the levels of pollution that will be generated each day. Since these models work by crunching data from from various atmospheric databases, they also calculate how wind currents will transport air pollution across any area at any point during the next days. However, accurate as they might be, these methods have various limitations. It takes many hours, and plenty of computational power, two forecast just two or three days ahead from the present moment. It is a complicated task even for supercomputers, and the annual growth of air pollution only means that calculating accurate forecasts will become more complicated as time goes by.

air pollution

A helping hand, or maybe two

At their best, traditional air quality forecasts can be accurate up to three days. More than than, and the models fail to pick on patterns, since computing is no longer powerful enough to comb through all of the data. However, Yunsoo Choi, an associate professor at the University of Huston, has been working on a new method that can forecast spikes in air pollution for as much as two weeks ahead of time. It does so by using Artificial Intelligence (AI).

Air pollution is a mixed bag of many chemicals. Known by most of us are Nitrogen Dioxide (NO2), and Nitrogen Monoxide (NO), as well as various PM 2.5 particles of different industrial origins. But Ozone (O3) is also part of this dangerous cocktail, even if we do not think about it as such. Atmospheric ozone makes up the outer layer that protects us by filtering out the Sun’s dangerous UV rays, but the story is different at ground level. Known as tropospheric, ground level ozone is created after any oxide of nitrogen (NOx) reacts to the presence of volatile organic compounds (VOC), such as benzene, ethylene and glycol. It is incredibly dangerous to our health, causing throat irritation, congestion and chest pain. It is one of the many chemicals that build up and pollute our cities, and professor Choi’s AI methodology was designed to forecast its concentrations.

A tool like this is very handy. If we can have a two-weeks certainty about the spikes of air pollution, and the type of pollutant that will be at the center of such concentrations, governments will have the opportunity to control the sources behind the emissions. Hospitals will also have a large enough window to prepare for possible admissions and decide how to staff emergency rooms, in case governments or policy makers are unable to do something about it. More than Artificial Intelligence, this model could be see as Assisted Intelligence, as the forecasts generated are good enough to help plan ahead of time.

Unlike traditional forecasting methods, professor Choi’s AI is trained by being fed previous air quality data. The AI then compares, studies, and models ozone’s behavioral patterns under various atmospheric conditions. Once the patterns have been discerned, an educated estimation is made about how much ozone will be present at any given moment within the next two weeks. The technology is very sophisticated and useful for all kinds of prevention and control, and it is just a matter of time before it is adapted to study, model and predict the concentrations of other air pollutants, like NO2 and black smoke.

air pollution

Benefits for all, but there is still work to be done

Although this is till a prototype technology, it is on the right track to be further incorporated into everyday practices. The data generate by professor Choi’s AI will be useful to all kinds of governing bodies, and hospitals will be the ones to obtain most of the advantages. Who knows, perhaps not far from the present moment we will see AI-powered air pollution forecast apps available for every smartphone imaginable. Tailored to our own needs, they may be able to warn us accurately about not being outdoors on any given day, or avoid certain zones at certain hours. Of course, this is just a vague idea at the moment, but it is well within the realm of what is presently possible.

Of course, there are still many things to do before professor Choi’s AI goes out the door and becomes a fully operational alternative to traditional air pollution forecasting. There are some limitations to the technology at the present time, and even some skepticism among the people involved with the project. For example, the AI might be giving right forecasts at the present time, but some of the researchers still don’t know if it is doing so due the right reasons. The data the system needs to comb through is vast and complex, with some really complicated math involved. Atmospheric science is subtle, yet chaotic, and professor Choi’s technology needs to be bullet proof before being green lit for mass implementation.

Then, of course, is the question of other pollutants. At the moment, the technology considers only spikes in ozone and its movements through a geographic area, but there are still plenty of other contaminants whose emissions and concentrations need to be accounted for. Air pollution forecast is not simple, but the technology is on the right track to make it at least more accurate and less cumbersome. The day will probably come when we will be able to simply look at our phones, see that the air will be particularly contaminated and decide to keep the windows in the office closed. That, however, doesn’t mean that indoor air pollution is any better for our health.

air pollution

That is why AIR8 is here.

Here at AIR8 we have developed our own line of filtering products to combat air pollution in the workplace. Our medical-grade HEPA-13 technology is effective against all forms of indoors concentrations of any kind of contaminant, such as tobacco smoke, PM 2.5 particles, NO2 and even viruses like Covid-19. They do so thanks to their combination of various stages of pre-filtration and filtration, such as Cold Catalyst and Activated Carbon.

Our filters are units that can be easily transported into your office at absolutely no cost. Once there, they can easily be moved from one room to the next and adjusted to any space’s demands. Low on energy consumption and as quiet as any machine can be, our products are silent enough to go unnoticed, so your employees won’t be disturbed by any noises.

You could even try our demo kit for one month for 50%, since our commitment to a cleaner, healthier, indoors air is true. We are backed up by the CE marking on all our products, which are made using only the best environmental practices approved by the European Commission. Unlike outdoors air, you won’t need an advanced AI system to tell you when your office’s air will be safe for breathing. With AIR8, we guarantee that your work space will have the cleanest, most sanitary, air possible. So why not give us a call and get a quote?

air pollution
Join our weekly newsletter
Contact us
Leave a message, or just your phone no. and we will call you back.
Close button
Close button