You are here

Société Nationale des Chemins de fer Français (SNCF)

SNCF is the historical French National Railway Operator.

SNCF in some numbers :

  • SNCF has 170,000 employees.

  • Every day, SNCF transports:

    • 904,000 travellers in Main and Regional Lines (244,000 for the only TGV trains)

    • 1,580,000 travellers in Paris area.

  • Every day, SNCF manages 14,000 trains :

  • 1,206 Main Lines trains (630 TGV)

  • 5,883 Regional trains

  • 4,753 Paris area trains

  • 1,765 Freight trains

  • Every year, SNCF transports :

  • 320 million travellers in Main Lines (87 million from TGV trains)

  • 572 million travellers in Paris area.

SNCF is one of the most important commuter-train operators in Europe. Each day millions of passengers are using SNCF trains. Their demand concerning security has never stop growing and is one of the 1st order priorities for SNCF as well as for the Public Transports Organisation Authorities.
Today, some trains have been equipped with video-surveillance equipments which allow to record an incident, and to recover the recordings when the train is stopped (often during the night following the incident). This video-surveillance systems do not allow following the incident in real-time, and then the incident is the most of the time treated too late… If the video-surveillance images could be recovered in real-time, without waste of time, SNCF would be able to treat the incident more rapidly, hence providing "in-time" appropriate crisis management (alert the concerned teams at SNCF and/or the Police…). The BOSS Project main issue is to develop a system capable to detect an incident and to send the associated video-surveillance data-flow to a control centre. For SNCF, three main technical problems have to be addressed in order to achieve the BOSS Project objectives:

  • Image acquisition and processing techniques: those techniques must be addressed in order to be able to correctly see what is happening on the train regardless of surrounding conditions (light, vibration, etc..).

  • Danger scenario recognition: scenarios must be developed to recognise potential dangerous situations. That is something harder to do in the confined area of a train coach than in the open space of a train station. BOSS will allow to increase performance and robustness of actual behaviour analysis systems. The robustness of behaviour recognition depends upon the accuracy of the low-level object tracking results. The expected improvements on tracking algorithms will, nevertheless, not always provide 100% accurate results. Behaviour analysis systems must reason with incomplete and/or missing information and therefore the reliability of detecting pre-defined suspicious behaviours varies. For example, detecting a passenger jumping over a ticket barrier could be detected with much more reliability than the aggressive behaviour of one passenger towards another. In every case, the reliability of detection of the correct behaviour could be significantly improved by further research.

  • • High-rate communications link between the train and the ground: as said above, the system must be able to transmit the video-surveillance data in real-time to control centre. At the moment, the only existent solutions allowing to implement a high-rate bi-directional link (the control centre must be able to remotely pilot the cameras and other equipments on the train) capable of transmitting a high quality video-flow from a vehicle are based on satellite communications techniques and need the use of big and expensive equipments. Those systems have been developed to deliver internet services onboard trains, and hence, they give more importance to the capacity of the downlink communication. But BOSS objectives concern most of all the capacity of the uplink communication… Moreover, satellite systems are for the moment "experimental" versions or prototypes, and are not able to achieve the required reliability needed by BOSS applications. BOSS will develop a terrestrial multi-technology system, more adapted to sub-urban trains (urban and sub-urban areas) than satellite systems, more reliable, and capable to achieve the required QoS needed by BOSS applications.

Role in the project

SNCF's Innovation and Research Department has been working on all of those problems during last years and can therefore bring its experience in those areas to the BOSS Project:

  • "Image acquisition and processing techniques" and "Danger scenario recognition" : SNCF's has treated this subjects in French National Project SAMSIT, and in European projects ISCAPS and TRIPS.

  • High-rate communications link between the train and the ground: those last years, SNCF has carried out many internal projects in order to test : the WIFI coverage of a train (ClicTGV Project), Icomera's hybrid satellite/terrestrial communications solution, 21Net's bi-directional satellite solution, WiFi train-ground connectivity at many different speeds (TGV Communicant Project)… At a European level, SNCF has tested the IST project FIFTH bi-directional satellite solution developed by Alenia. Also at a European level, SNCF participates to the project SAET, of the European Space Agency. At French National level, SNCF participates to the project TRAINIPSAT (with INRETS and THALES).

Contact details(for authenticated users only)