According to the World Health Organization (WHO), air pollution in the last two decades alone has led to a reduction in life expectancy worldwide by more than two years, compared with the fact that air quality met the established standards. To determine the air quality, the air quality index or AQI is used - a relative value that can take values ​​from 0 to 500. The higher the value AQI, the higher the level of air pollution. AQI calculated on the basis of data on the concentration of sulfur dioxide in the air (SO2), nitrogen dioxide (DO NOT2), carbon monoxide (CO), ozone (O3) and suspended solids (PM10). In addition to these substances, volatile organic and inorganic compounds (such as ammonia, formaldehyde, nicotine, benzopyrene, toluene, xylene, chlorine, their derivatives, and many others) are also in the air. Thus, the assessment of air quality is impossible without assessing the concentration of all types of pollutants in its composition.

Existing solutions in the field of air quality assessment are mainly intended for the analysis of the concentration of any one pollutant and use only one type of gas sensors as a sensing element. This significantly limits the functionality of the system. For example, if a system for determining air quality is built on the basis of electrochemical gas sensors, then its capabilities will be limited to measuring concentrations only  DO NOT2, SO2, O3, DO NOT... At the same time, the measurement result of electrochemical sensors is strongly influenced by changes in temperature and humidity. To determine the concentration of volatile organic compounds, the use of photoionization detectors is necessary. However, this type of sensor can only detect substances whose molecules are ionized by ultraviolet radiation. In addition, all existing solutions have a number of disadvantages:

·        lack of scalability c integration into the management system of a city or region (building a network of systems for analyzing air quality in an entire region or several remote ones);

·        are often just an idea, and not a really developed and mass-produced device;

·        work according to the classical algorithm (they do not have a technique for detecting and registering unknown polluting compounds);

·        the problem of non-linearity, low selectivity and cross-sensitivity to other gases;

·        scatter of readings of the same type of sensors due to different degrees of degradation of the sensitive element in different territorial sectors.

Thus, an effective assessment of air quality is possible only when using modern systems for detecting pollutants using a combination of several types of sensors and state-of-the-art intelligent data processing methods recorded by these sensors. The main purpose of such systems is to detect all known and potentially hazardous gases in the environment and measure accurate readings of their concentration, and then ensure appropriate actions and measures to ensure the unconditional safety of humans and the environment. Hazardous gases are detected using gas sensor arrays, and fuzzy logic identifies two adjacent (or derivative) and unknown gases. Using fuzzy logic classification in an electronic multi-range detector system, hazardous levels of known and unknown pollutant gases can be determined.

Current gas sensors have problems with non-linearity, low selectivity and cross-sensitivity to other gases that cause significant deviations from expected results.

Using a modern approach using hardware control based on an artificial neural network, it is possible to identify complex gas mixtures with a multi-zone classification. The fuzzy logic model for the gas sensor array will effectively detect the presence of unknown gas contaminants.

A simple explanation of fuzzy set theory	A simple explanation of fuzzy set theory

The use of an artificial neural network model and deep learning technology will provide an opportunity not only to efficiently assess in real time the degree of pollution and the expected pollution of the environment, but also to suggest the presence of unknown pollutants or their mixtures in the atmosphere. This will make it possible to predict changes in the ecological situation, taking into account weather conditions (temperature, humidity, wind speed and direction), and, as a consequence, predict conditions. And as you know, forewarned is forearmed.

Air quality assessment systems built on the basis of such solutions are integrated into a single network of an enterprise (group of enterprises), city, region and country as a whole, making it possible to assess the degree of air pollution and form interactive maps of the gradient of the ecological situation.