The project uses real-time data from partners in Southern California at the city, county and state level. Credit: Alexander Schimmeck on Unsplash.

A team of US computer scientists from Lawrence Berkeley National Lab and the California Department of Transportation (Caltrans) is exploring high-performance computing (HPC) and machine learning for real-time traffic analysis.

The research was carried out in conjunction with California Partners for Advanced Transportation Technology (PATH) and Connected Corridors.

The system is being implemented by Caltrans and Connected Corridors on a trial basis in Los Angeles County through the I-210 pilot.

Berkeley Lab Computational Research Division (CRD) mathematician Sherry Li said: “Many traffic-flow prediction methods exist, and each can be advantageous in the right situation.

“To alleviate the pain of relying on human operators who sometimes blindly trust one particular model, our goal was to integrate multiple models that produce more stable and accurate traffic predictions. We did this by designing an ensemble-learning algorithm that combines different sub-models.”

The project uses real-time data from partners in Southern California at city, county and state levels to improve Caltrans’ real-time decision-making.

Next year, the initial iteration of the system will be deployed in the California cities of Arcadia, Duarte, Monrovia and Pasadena.

Project funding was provided by Berkeley Lab’s Laboratory Directed Research and Development (LDRD) programme.

As part of the project, a computational framework will be built to enable HPC applications specific to transportation, such as optimisation and control of traffic equilibrium.

The project used data collected from Caltrans’ Californian highway sensors and generated new algorithms that achieved accurate prediction on a 15-minute rolling basis.

Later on, the algorithms were validated and integrated using real-time traffic data collected from the Connected Corridors system.

The project aims to generate predicted traffic flows at points where sensing is present on the freeway.

The research team applied real-time learning techniques to enable the algorithm to learn from the past and adapt to new traffic conditions in real-time.

For accurate and timely traffic prediction and to aid real-time traffic control, the algorithm could be used in combination with these technologies.