Pierre Delaigue: “Autonomous mobility will help take down the walls around non-urban territories”

Automated and on-demand transport will play a definite part in the mobility services of the future in rural and sparsely populated areas. That is, provided the infrastructure fulfils its function, a task which Pierre Delaigue, Director of Autonomous, Connected and Electric Mobility Projects at Leonard, has taken on in collaboration with several of the operating entities at VINCI Group. An interview.

“Autonomous vehicles could prove essential in reconnecting the countryside and designing mobility solutions for rural dwellers” one leading firm estimated in 2018. Are you of this same opinion?

Absolutely. The autonomous mobility sector has been developing for several years now, mainly on motorways and in cities. However, it is important that these two environments not overshadow what is known as the “secondary network”. Until now, sparsely populated areas have also been the most largely neglected when it comes to autonomous mobility.

The offer we envision would contribute to taking down the walls around specific territories where there is no longer any mobility offer (for example, rail or bus lines that are no longer in service) or an inadequate one. Autonomous mobility, with a modular capacity that would be adjusted to demand, can replace offers that society has deemed too costly to sustain. It can also extend the service ranges of existing mobility offers, for example, when bus lines are still running, but at insufficient frequency or time slots.

 

How can such solutions be developed?

We are working to facilitate the deployment of stand-alone solutions in low-density environments. Experiments are being set up, and the idea is to move forward in a partnership approach, with vehicle suppliers, in a spirit of co-development.

Autonomous mobility is traditionally a chicken and egg problem: either you start equipping the infrastructure until such time as compatible vehicles are available, or you wait for a sufficient critical mass of vehicles, then equip the infrastructure. It is our belief that it is better to work hand in hand with the manufacturers and be in line with the roadmaps for the arrival of autonomous vehicles. Our approach is incremental: as needs are identified, we offer new solutions.

When it comes to autonomous and connected mobility, we are thinking about tomorrow’s mobility services, but we are laying the foundations today, by mobilising an e-ecosystem and building a development timeline.

 

Is the rural environment also fertile ground for experimentation?

Motorways are now a privileged ground, facilitating testing and development, as they are highly controlled linear routes, with uniform speed ranges, reduced vehicle diversity, a protected environment, and no intersections…

Other environments, whether urban or rural, offer their share of technical challenges that must be addressed. In the city, these include: density, a wide range of moving objects, frequent obstructions and signal transmissions that are disrupted by buildings. In rural areas, while traffic density is less important, the vegetation, which moreover changes according to the seasons, creates a large number of parameters that have to be dealt with. Roadside equipment, including seldom present markings, also requires attention, as it is crucial.

 

 

On motorways or on the secondary road system, what role should infrastructure play in facilitating the development of autonomous mobility? What are the conditions for such deployment to take place?

There are several levels of infrastructure support for autonomous vehicles, first of all road equipment, i.e. markings, signs, barriers, etc., [see Eurovia’s work on this subject, Editor’s note]

Another building block is connectivity and sending information using multiple types of existing technology, both short and long range. We have no particular religion on connectivity technologies and will adapt to market developments.

Another building block is that formed by so-called “off-loaded” data, and in particular roadside perception: to provide autonomous and connected vehicles with the most qualified and real-time information possible on traffic conditions (detecting incidents, characterising flows, determining the position, speed and dynamic behaviour of each of the objects moving in the lane), we call on the skills of Cyclope.ai. A company run by Vinci Autoroutes, Cyclope.ai develops artificial intelligence solutions to serve road and motorway infrastructures. These use machine learning to interpret images from roadside cameras, and thereby provide vehicles with a broader understanding than that offered by their on-board sensors alone. With these sensors, the vehicle sees only 200 m away, i.e., on motorways, can project only 5 seconds ahead; here, we enable them to see beyond this on-board horizon, giving them extended perception. To do this, we use equipment (cameras and cellular networks) already found roadside. Whenever we feel this is not enough, in our experiments we add local connectivity points, in addition to the long range network, in order to have full control over the connectivity at specific “singular points of the network” (toll station, slip road, etc.).

Lastly, as concerns data platforms, we already have information systems on our motorway network that centralise data on traffic conditions. VINCI Autoroutes has been working for years on digitising these IS. This preliminary work will make it possible to accommodate potential changes in the future around the needs of autonomous and connected vehicles.

 

Is infrastructure ultimately a critical component in the deployment of autonomous mobility?

There is currently growing consensus that infrastructure support is undoubtedly of great value for the autonomous and connected vehicle. While different levels of Infrastructure-To-Vehicle (I2V) support remain possible, the experiments being carried out now will enable us to specify exactly how infrastructure contributes depending on the road environments, territories and mobility services targeted.

Share this article on