carpooling logo par

Dedicated lane control system,
to support carpool lanes deployment has developed a robust software solution that can detect the number of occupants into a vehicle travelling on restricted lanes

contact us

A solution to better share the road

The French Mobility Policy Act (LOM law), adopted last year, prescribed the introduction of dedicated carpool lanes as one of the key measures to meet the ecological challenges of the Low Carbon Highway and relieve congestion on the road network. To support this initiative, has developed a video analysis solution that can count the number of occupants in a vehicle travelling on the road.

Thanks to state-of-the-art deep learning technologies and a significant investment in R&D, our solution detects automatically the number of occupants present in the vehicle : for front and back seats, both day and night.

Ask for a demo

Available technological functions

Counting the number of occupants

Lateral and frontal detection for an accurate counting

Vehicles classification

Vehicle class recognition according
to pre-defined categorizations

License plate reading

Linking of vehicle information with the license plate number

Why Carpooling?

High detection rates

A very high level of performance achieved, thanks to a cutting-edge expertise and several preliminary models trainings

Minimized impact on operations

Operational interventions optimized thanks to a compact solution on the roadside

Easy to deploy

A modular, plug-and-play, low intrusive system to facilitate installation and maintenance

Compliant with privacy rules

Guaranteed thanks to a in-house anonymization technology: systematic blurring of faces to protect personal data

Ask for demo

Ongoing experimentation at VINCI Autoroutes

VINCI Autoroutes Logo

Lançon-de-Provence, A7 highway




The aim of this experiment is to improve and qualify the performance of high-speed detection models (+90km/h), while identifying the optimal set-up and configuration for dedicated lane monitoring system from the roadside.


To support the R&D and design phase of the Carpooling product, VINCI Autoutes (ASF network) provided with an experimentation area for the installation of a solution prototype equipped with camera, in order to improve the system algorithms.

CEREMA EVALUATION has mandated CEREMA to evaluate by this summer its firt device performance directly on the test site.

Personal data Processing

As part of this experiment, is responsible for
the collection and processing of personal data of A7-motorway users:

Treatment details:

Collected data

Detection of silhouettes and plate numbers: the occupants’ faces are systematically are irreversibly blurred.

Data retention period

All data collected and then anonymized (faces blurring) are deleted after 30 days.

Right of access

For any request to exercise their right of access, rectification, opposition, limitation, and deletion of personal data, users can write to

Legal framework

These processing are carried out in compliance with Law No. 78-17 of 6 January 1978 on Data Processing, Data Files and Individual Liberties, amended by the (EU) Regulation 2016/679 of 27 April 2016 on the protection of personal data (GDPR).

Users information

A7-highway users are informed upstream of the experimentation site by a road sign indicating "Carpooling sensor test".

At the service of mobility management authorities

As part of its R&D phase, wants to strenghthen its algorithms in order to have a product that delivers a very high level of quality. The Carpooling solution must meet the needs of road infrastructure operators and other mobility mangement authorities for a better control of carpool lanes uses.

Collection of "good" images

The first step for the proper development of an efficient monitoring device is to ensure qualitative data collection: the quality of the image that is going to be processed by the algorithm has a key impact on the final overall system performance.

Contcat us

A compact roadside device

Our product meets all the functionalities of a control device, while respecting the need to optimise interventions during operations: compact device, protected equipments, easy to install, accurate site-specific fine tuning to adapt to each specific set-up, etc.

Learn more

Performance acceleration

Thanks to deep learning technologies, our algorithms are constantly improving. Indeed, the more data they “see”, the more accurate the models are, by learning how to recognize previously undetected cases.

Learn more

At-scale deployment & evaluation tool

In order to evaluate the performance on each site in an accurate way, provides the operators with a dedicated tool, that can analyze and qualify all our AI models. Once the required level of performance has been reached, the Carpooling algorithm can then be safely deployed at scale, direcly on the network.

Contact us

Advantageous political context

The French Mobility Policy Act (2019), amended the existing legislative framework for bus and taxi lanes, allowing several MOBILITY MANAGEMENT public AUTHORITIES (regions, metropolises and local communities) to extend these lanes to vehicles occupied by at least 2 persons (VR2+)

Contact us

Product functionalities


Analysis tool provision to guide the operator in the solution at-scale deployment, through performance monitoring at each step


Possibility to put additional functionalities: classification based on vehicle types (buses, taxis, priority vehicles), license plate reading, etc.


The API returns a complete vehicle record in less than 500 ms, allowing real-time monitoring


A confidence rate is associated with each response, working as a real filter to guarantee a predefined level of reliability

Carpooling Interface