The AI Research Center at Novosibirsk State University has completed the development of algorithms for monitoring atmospheric air quality. The project was commissioned by CityAir, a company specializing in environmental solutions. The new software improves the accuracy of sensor data processing while reducing the computational requirements for the hardware.
The research team, led by senior researcher Vladislav Perepelkin, presented the results at a new monitoring station in the Oktyabrsky district of Novosibirsk. The station is located at the Novosibirsk State University of Architecture and Civil Engineering and was established in collaboration with the regional Ministry of Natural Resources and Ecology.
The developed adaptive algorithms account for external environmental factors, such as temperature and humidity, which typically affect sensor accuracy. By analyzing historical data, the AI compensates for sensor degradation over time, reducing the need for frequent manual recalibration or equipment replacement.
The project also focused on lowering the computational complexity of data analysis. By replacing traditional, resource-heavy methods with AI-driven approaches, the system can now run on compact hardware. This allows for the installation of monitoring devices on urban infrastructure, such as utility poles, without the need for dedicated facilities.
For the university's AI center, this project marks a transition from theoretical research to practical industrial applications. The team plans to leverage the experience gained to develop similar monitoring solutions for other urban networks, tailoring the algorithms to specific equipment and operating conditions.
