AI-based approach to fast detection and resolution of urban traffic network bottlenecks

It is well known that the road networks are congested in most cities. What is less well known is that the capabilities of these road networks, however, are far from being fully utilized, and an optimized network solution can reduce real transport delays by at least 5%, and sometimes by up to 25%. To find such a solution, we use spatial multivariate analysis in conjunction with an advanced TOD traffic control system

In the course of many years of research, we were able to create an automated system for automated determination of travel delays on urban roads with high resolution (down to 35-ft segments). Analysis of travel delay arrays in conjunction with highly accurate Ticon traffic volume estimations allows traffic engineer to conduct bottleneck ranking and promptly provide recommendations for mobility improvement.

In 2023, we conducted field tests of this approach on a suburban network of a metropolis serving a population of 200,000 people (Irpin city agglomeration, Ukraine). We analyzed the results of these trials in relation to large cities in the United States (using the examples of Atlanta, GA; Boston, MA; Reno, NV; as well as five5 smaller cities), and obtained results that instill hope for a significant (on average 7-10%) increase in the urban mobility by implementing our methodology.

This result becomes even more promising, as the method leverages only AI-based optimization solutions and does not require the purchase and installation of a new hardware.