Supercomputer Enhances the M1-M7 Smart Highway – A New Era of Real-Time Data Analysis Begins

A new supercomputer has arrived for the experimental smart section of the M1-M7 highway, enabling real-time data analysis. The BME Automated Drive team is working on integrating the system to provide accurate, instant information based on collected data, making traffic safer and more efficient.

The M1-M7 highway joint introductory section has taken another significant step forward in traffic innovation through its smart highway project. While sensor calibration is ongoing, the arrival of the supercomputer, capable of processing vast amounts of data in real time, represents a major advancement in the project.

András Rövid Ph.D, the technical lead at BME for the Eureka Central System project, stated that the supercomputer plays a critical role in extracting accurate, real-time traffic data. “This device provides high computing capacity, allowing AI-based data processing and prediction in real time. We can instantly track crucial details, including the precise position, size, and movement direction of vehicles,” Dr. Rövid explained.

The sensor network on the M1-M7 section generates an enormous amount of data. A total of 12 LiDAR sensors, 21 cameras, and 6 RADAR units monitor traffic, generating up to 1.5 Gigabytes of data per second, equivalent to 130 terabytes daily. The camera array includes narrow and wide-angle cameras, thermal cameras, and fisheye lenses, providing comprehensive, multi-directional views of the traffic situation. This complex system generates up to 20 frames per second, which the Edge computer must process in real time.

The computer has remarkable capabilities: it includes two Nvidia H100 GPUs, each with 80 GB of memory, over three times the memory of a high-end consumer graphics card, like the popular 4090 GTX model with “only” 24 GB of memory. In total, 160 GB of memory is available, providing substantial capacity for complex calculations, such as determining vehicle positions, movement, and size from camera, radar, and lidar data. The so-called Edge computer is equipped with over a terabyte of memory and significant storage capacity. It can reach up to 3 kilowatts of peak power while running various algorithms to analyze and process road data.

“Real-time processing is crucial, as we may need to analyze data from hundreds of vehicles simultaneously within 100 milliseconds during peak traffic,” András Rövid explained.

The new supercomputer will soon be transferred to its final location at Magyar Közút’s data center in Szigetszentmiklós, where testing will begin. The BME Automated Drive team will continue refining the algorithms to ensure the system can quickly and accurately interpret and predict traffic conditions.

One of the smart highway’s goals is to provide real-time, vehicle-specific information to drivers—even those without sensor-equipped vehicles—through mobile applications, offering warnings about the position of cars ahead, for example. This improves traffic safety and helps streamline traffic flow. This technology is currently unique in the region and is a vital step in enhancing Hungary’s digital infrastructure. The M1-M7 demonstration section not only improves traffic safety but also lays the foundation for the future of transportation.