The unbounded potential of Artificial Intelligence: the case Périphérique Nord of Lyon – SE BPNL

Sprinx's innovative approach has allowed the revamping of the traffic analysis and video surveillance system of the Périphérique Nord in the Lyon metropolitan area.


The Lyon region is a historical crossroads of trade and transit, not only within the French territory but also from an international point of view, due to its proximity to the significant European tunnels, Frejus and Mont Blanc.

The lack of a hierarchical ordering of the roads in the urban area has generated difficulties in traffic for several years, especially during rush hours. For this reason, in the north of the Lyon metropolitan area, the Boulevard Périphérique Nord was born in 1997, a complex urban toll structure with a total length of 10 km, a junction point with the A6 (Paris / Marseille), A42 (Geneva) and A43 (Grenoble / Chambéry). With its 4 tunnels for a total of 6.5 km in length, 2 viaducts, 7 gates, 2 toll booths, the BPNL has established itself as one of the most strategic infrastructural works, improving the quality of mobility and urban viability.

In 2020, the SE BPNL (Société d’Exploitation du Boulevard Périphérique Nord De Lyon), assisted by HGM Ingénierie, launched the complete revamping of BPNL’s existing traffic analysis and video surveillance system.

The challenge involved updating about 200 cameras throughout the section, both outdoors and inside the tunnels, and the respective video recording and analysis systems. This was a very ambitious project as regards the Automatic Incident Detection system.

Sprinx has won the contract thanks to the consolidated experience and innovation presented in the project.

The challenge
The Concessionaire’s request was to reduce service suspension times to a minimum during updating existing systems, minimizing inconvenience to city traffic and operators and the indirect and direct costs due to the blocking of transits. This intervention had to ensure, even in the preparatory stages for the transition to the new one, that there were no reductions in traffic monitoring performance compared to those of the system previously in operation.

The solution
The Sprinx and Bouygues Energies & Services consortium proposed an ambitious project to limit downtime, minimize disruption, and ensure constant growth in traffic monitoring performance. The project consisted of two main crucial phases, as requested by the Contracting Authority, and as regards the AID system, we proceeded as follows:

  1. First Phase: Parallel activation (pré-basculement) to the Sprinx AID system’s existing system, using the video streams of the existing cameras. The parallel activity made it possible to demonstrate that Sprinx’s platform could guarantee, even with obsolete field equipment, good performance compared to that of the system in use by the Dealer. Thanks to the innovative combination of Deep Learning and 3D object tracking, is, in fact, an incredibly adaptive platform, which guarantees high detection performance minimizing the number of false alarms even under challenging environments. Furthermore, the combination of Sprinx’s ten-year experience in vehicular and pedestrian mobility analysis with the most modern technologies based on Artificial Intelligence has made it possible to integrate extremely high-performance and user-friendly configuration, calibration, and tuning tools into the platform, allowing easy and fast deployment of a technology based on complex and sophisticated algorithms.
  2. Second Phase: The pre-existing analysis system was definitively shut down and replaced with Then we proceeded with updating the field equipment: the replacement of all existing cameras with new generation IP cameras. The activity took place gradually at night on all internal and external cameras to minimize service blocks. Thanks to Sprinx technology, it has been possible to optimize the reconfiguration times of the new cameras on the system, guaranteeing a migration to the new system in an almost transparent manner for the operations room operators. The increase in the traffic monitoring service was also certified by the deep level of integration ensured by the Sprinx plug-ins with the Milestone Systems video recording and surveillance system, unique in the sector, and by the Sprinx OPC-UA out-put modules with the existing radio room ITS system. The unified management and visualization platform of the systems, Sprinx dragon, was also fundamental. It allowed SE BPNL operators, through an easy and intuitive web interface, even in the migration phases and after the activities, to constantly monitor the progress of the AID system, querying the event log and viewing the graphs relating to traffic and system alarms.

Sprinx solutions added value: AID system and dragon management platform

The combination of artificial intelligence algorithms and the 3D object tracking technology for an easy and performing calibration and configuration made it possible to make the implementation and activation process of the new AID system fast and precise, ensuring a maximum reduction in downtime requested by the Concessionaire.

Furthermore, Sprinx was the only company able to offer, within the project, deep integration with the Milestone video recording and video surveillance system, thanks to the Plug-ins and Add-ons developed. In fact, through the Milestone client, BPNL radio room operators can access and view all AID information (alarm events, traffic statistics) generated by the Sprinx system in a fully integrated way with the video-management system.

Sprinx dragon platform, then, not only was a unified interface for displaying and managing all the AID systems for both operational and maintenance purposes but also a communication gateway with the radio room ITS system, capable of allowing not a simple interface but a deep integration between commands and system logics, ensuring a more robust and reliable architecture.

The results
In a short time, the results achieved were:

  1. Use of AI technology: Modernizing the system, replaced with the latest generation one (AI technology), based on standard PC platforms with Intel® OpenVINO TM processor.
  2. Time optimization: minimizing the revamping times of the existing system with a new generation system perfectly integrated with the video recording system and with the radio room ITS system.
  3. Enhancement of open-air detections: the application of traffic analysis algorithms in outdoor environments has always been affected by perturbations generated by changes in light and atmospheric conditions. The technology proposed by Sprinx, 3D object tracking and Deep Learning module, has made it possible not only to guarantee high performance also for the detections requested by SE BPNL in the open air but to add others thanks to the results demonstrated.
  4. Project with additional potential: in addition to the AID with detection of the main events in the open air and the tunnels, Sprinx has integrated the activation of statistical functions for counting and classification of vehicles, which is useful in its proposal for predicting traffic and count. This will allow SE BPNL to exploit the video surveillance cameras to collect statistical data, evaluating the replacement of the sensors currently in charge of these functions (e.g., radar), thus allowing further management and maintenance costs optimization. Furthermore, since it is an AID application but in an urban context, functions developed by Sprinx for the world of Smart Cities, such as the detection and counting of bicycles or scooters, can be easily integrated, ensuring the customer a continuous growth of the system.

Source: EDI-CEM Srl – Rivista “Strade & Autostrade”

To find the Italian version –