Researchers from the Texas Advanced Computing Center (TACC), University of Texas for Transportation Research and the City of Austin in the US are developing a new artificial intelligence (AI) tool that will enable searchable traffic analyses to resolve traffic issues in Austin.

The new tool will use raw traffic camera footage from the cameras installed at the streets of Austin to identify moving objects such as pedestrians, vehicles, and bicycles, as well as traffic lights.

It will obtain information on how these objects move and interact in the streets.

Traffic engineers will analyse the information to determine specific queries and devise algorithms.

TACC Data Mining & Statistics Group research scientist Weijia Xu said: “We are hoping to develop a flexible and efficient system to aid traffic researchers and decision-makers for dynamic, real-life analysis needs.

“We are hoping to develop a flexible and efficient system to aid traffic researchers and decision-makers for dynamic, real-life analysis needs.”

“We don’t want to build a turn-key solution for a single, specific problem. We want to explore means that may be helpful for a number of analytical needs, even those that may pop up in the future.”

The algorithm developed for traffic analysis automatically labels all potential objects from the raw data and tracks them with previously recognised similar objects to compare the output.

After the system was capable of labelling, tracking and analysing traffic, it was used in real-life situations such as counting the number of vehicles for a specific time.

Preliminary results show that the system was 95% accurate.

It is expected to help in developing transportation models on the basis of traffic volumes.

The system can also identify close encounters between vehicles and pedestrians and help to detect risky locations.