bg
News
21:27, 07 March 2026
views
9

AI System in Russia Is to Assess Road Conditions and Plan Repairs

Artificial intelligence is expected to help adjust maintenance schedules based on the actual condition of each road.

Photo: iStock

Researchers at Don State Technical University are developing an artificial intelligence system to assess road conditions, professor and acting dean of the university’s Road Transport Faculty Artem Tiraturyan told TASS.

According to him, the work is being carried out in several directions. One of them involves creating optimal road maintenance schedules using evolutionary algorithms. Another focuses on applying AI to monitor pavement conditions using incoming diagnostic data.

“We plan not simply to follow a fixed schedule—routine repairs every 12 years and major overhauls every 24 years—but to adapt maintenance work to the actual condition of each road. In some places repairs may not be necessary by the scheduled time, while elsewhere deterioration may progress faster and require more frequent work to prevent major financial costs,” Tiraturyan told TASS.

A Diagnostics-Based Management Concept

The professor added that the university plans to develop its own lifecycle management concept for road infrastructure based on diagnostic data collected in previous years.

The system is expected not only to forecast changes in road conditions and help plan repairs more efficiently but also to improve how maintenance funds are spent.

Earlier reports noted that a neural network used to monitor roads in the Moscow region reduced the number of complaints by 10 percent. Mobile diagnostic complexes equipped with an AI-based application are installed on service vehicles.

The built-in neural network algorithm analyzes every meter of road surface in real time. The tool detects potholes, pavement defects, debris, and other violations. The information, including precise geographic coordinates, is sent to contractors responsible for fixing the issues.


like
heart
fun
wow
sad
angry
Latest news
Important
Recommended
previous
next