[Interview] Autonomous vehicles: reality is not, and will not be, what you think…
Autonomous vehicles are the subject of so many announcements that it is hard to know what to believe. An automotive journalist for more than 30 years, Laurent Meillaud keeps a close watch on technological developments in the sector, and the challenges facing it. Today, he gives us his take on what’s already out there in terms of autonomous vehicles, what might emerge, and what will probably remain in the realms of fiction…
To define what we are talking about, can you explain what is meant by autonomous vehicle?
It is a vehicle capable of moving and maneuvering without any driver aboard. This is the highest level of automation in the SAE classification (Society of Automotive Engineers). In this classification, level 0 is completely un-automated vehicles and level 5 is vehicles with fully automated driving, capable as a consequence of driving without human intervention on all types of roads, in all conditions and anywhere in the world. This classification applies to all vehicle categories: private cars, public transport and utility vehicles (HGVs, LUVs, construction machinery…)
Let us be clear: currently there are no level 5 vehicles. And for both technical and regulatory reasons it is possible that this level will never be achieved.
Private cars coming off the production line today are level 2, with driver assistance systems that enable the driver to take his hands off the wheel from time to time, for example in traffic jams. On the other hand, we have achieved level 4 on certain types of automated shuttles/minibuses. They have been in trials for years, typically in a protected site or on private premises, but it is now known how to operate this type of shuttle in regular traffic – at reduced speed and subject to certain modifications. A public transport operator such as Transdev has been working in this space for more than 10 years now. It can boast millions of kilometers driven and hundreds of thousands of people carried using these shuttles, throughout the world, and with no personal injury.
True, this is not level 5, because these shuttles only operate on well-defined routes, but it is a very interesting option to add to a public transport offering. Moreover, this is the government’s preferred approach, especially in France where the Framework Act on Mobility enables and facilitates the deployment of highly automated vehicles in locations where the only other option is the private car.
What credence do you give to the scenario in which private cars are replaced by fleets of autonomous vehicles criss-crossing our towns and cities and transporting people on demand?
This might be possible in 20 or 30 years. From an environmental perspective, the scenario makes sense if you consider that private cars are currently stationary more than 9/10 of the time, and therefore very much underused.In 2015, the International Transport Forum (OCDE) modelled this scenario for the city of Lisbon. This study shows that deploying a fleet of 6000 autonomous taxis operating around the clock could reduce the number of cars on the road in this city by 90%, without significantly affecting its inhabitants’ mobility in terms of journey start point, destination and travel time. This scenario eliminates bottlenecks while also freeing up 95% of parking places for other uses. But this remains a simulation and nowhere in the world is there any question of completely replacing the private car with shuttle fleets. There are of course already robotic taxis in China, but even if the Chinese are forcing the pace this remains confidential and experimental.
The medium-term reality will be less spectacular than what we have been sold! In my opinion, we will an autonomous shuttle offering as a public transport add-on and, in parallel, the development of car-sharing systems between individuals, a more sophisticated version of what currently exists because these cars will be both far more automated and connected.
Are we further on with free transport? What about autonomous HGVs?
All the manufacturers are working on this issue. There is a real will to make it happen, both on their part and by the authorities. With the increase in volumes carried by road and the explosion in e-commerce-related deliveries, autonomous freight transport does indeed look like an interesting solution in two respects: on the one hand for the environment and the climate, because autonomous truck implies electric truck and therefore “zero emissions”; on the other hand to make up for the driver shortage confronting the road transport industry. But here too, despite numerous announcements and trials throughout the world, there is still no roll-out and the number of electric or hybrid trucks on the road is derisory.
The ideal for long haul transport would be to have motorways or at least motorway lanes dedicated to truck convoys. There were projects along these lines in France, but they have been abandoned. Germany has been testing a 5 km stretch since 2019 where electric trucks recharge from an overhead catenary, like trams. But for the time being these are not autonomous trucks, and this type of infrastructure requires enormous investment! The truth is that mines, quarries, ports and airports are currently the only places where trucks and other autonomous vehicles are truly operational, albeit still remotely operated.
What possible scenarios are there for last mile delivery?
Some manufacturers, but start-ups as well, have addressed this issue and the Covid crisis has accelerated the trialing and deployment of new solutions. In China and the United States, we have seen autonomous vehicles and robots making deliveries in towns and on campuses. Technically, we know how to do it and there are several interesting concepts. For example, you can have highly automated, if not autonomous, vehicles from which operate robots that carry the parcels to the point of delivery, autonomously or by following the delivery person on the pavement. It is a valid option for reducing the number of delivery vans in dense urban or pedestrian areas, but also for technicians or tradespeople who need to carry material to their place of work. For the past two years, TwinswHeel droids made by the French company Soben have been accompanying Enedis technicians through the streets of Toulouse. It is fully operational.
10 years from now, rather than a driverless delivery vehicle moving around on the road all by itself, I see the emergence of a highly automated utility vehicle actively assisting the driver, which parks itself, thereby enabling him to concentrate on other aspects of his route, for example receiving calls or calling customers before his arrival. The vehicle’s advanced automation will make driver-delivery personnel’s work less arduous and keep them safer by eliminating excess speeds and safety hazards. Moreover, this vehicle will of necessity be connected, it will be “intelligent” and capable of changing the sequence of its route in real time depending on the traffic, or to incorporate an additional stopping point, or a cancelled appointment.
What grounds are there for saying that these utility vehicles and delivery solutions may become ubiquitous in the next decade?
We can already do a great deal and the pace of innovation is very rapid. When it comes to automation, this starts with the top of the range. But nowadays the vehicle manufacturers rapidly roll out anything to do with driver assistance. Light utility vehicles, hitherto seen as the poor relations of the automotive industry, are also benefiting from reversing cameras, sensors and other driving aids. As this is something that is developing at warp speed on trucks, utility vehicles will also benefit, and all the more quickly as these developments address a genuine need by professionals as well as local authorities’ expectations.
What are the main obstacles to the deployment of genuinely autonomous vehicles?
What you have to understand is that no autonomous vehicle is possible without ubiquitous network connectivity, within the vehicle and with the outside. It is an indispensable prerequisite for ensuring the transmission of signals and to be able to take control of the vehicle remotely, which is mandatory under current legislation. A connected vehicle generates a phenomenal volume of data which needs to be captured and processed, especially if equipped with numerous sensors, as is the case with autonomous cars. Existing networks are inadequate. 5G, deployment of which is in its infancy in France, will provide the bandwidth and extremely small latency period that will enable these remote operations, as well as communication between the vehicle and other vehicles, and the environment.
Next up, even if you are thinking about roads, lanes or areas reserved for autonomous vehicles, specific adaptations and equipment are required – if only connected lights which give them priority over other vehicles, or vice versa, or else cameras to identify the vehicles that are permitted to use these lanes. The equipment cost or cost of upgrading road infrastructures to make them “intelligent” will be extremely high, which precludes any general deployment, even in the medium term.
To enable a genuinely autonomous vehicle to drive on the road also requires ultra-high definition maps updated in real time, far more sophisticated than what we have in our smart phone apps. In order for the autonomous vehicle to drive safely, accuracy needs to be centimetric, and any change, even temporary, must be capable of being reported and incorporated instantly. This assumes an environment permanently monitored by cameras and sensors which are continually uploading information on traffic, the state of the roads, roadworks, etc. This cartographic offering is still very limited, but actors such as HERE and TomTom are working on it.
Finally, there is the cost of the vehicles themselves. Autonomous vehicles are equipped with very expensive sensors and data technologies. But even if they have these technologies, this is not where the companies in the sector are going to focus their efforts today. The priority is electrification: to move towards “zero emissions” and to comply with targets in Europe and in other countries, they have no choice but to transition to electric motors and this too is a very heavy investment.
Is Europe lagging behind in these areas?
If the autonomous vehicle is to become everyday reality, enormous public and private investment needs to be mobilized and coordinated over a sustained period. Which is why I do not believe in the autonomous car becoming ubiquitous; instead, I believe in local and add-on solutions such as automated shuttles, sophisticated car-sharing platforms and the delivery systems I spoke about earlier.
The autonomous vehicle per se is not a priority for European agencies and states but this subject unquestionably encompasses major economic and sovereignty challenges in an industry which is global in any event. European companies are being challenged by American and Chinese groups from the digital sector who want to pre-empt the autonomous vehicle market and it would be a good idea to protect ourselves from them. Here in Europe, we have companies that are global contenders. Automotive supplier Valeo, for example, is among the best.
But for reasons that are beyond my comprehension many people seemingly revel in saying that everything comes from Silicon Valley, from Google, from Apple and from Tesla. Although Google has been up to things for the past 10 years and has been committing resources, we have never seen the merest hint of a prototype autonomous car at Apple! Tesla is an enigma: although it was indeed the first to do remote updates, its automated driving system is no more sophisticated than what you will find in top of the range cars from “conventional” manufacturers. The self-parking car existed before there was a Tesla! As for inter-vehicle communication, Tesla is doing nothing and never talks about it! You would do better to look to China, which has made the autonomous electric car a national challenge and which, with very strong political support, is creating champions in this space.
Be that as it may, the big Western manufacturers are not remotely behind the curve, they are all on the case, are making electric motors, and there are several start-ups providing technological solutions that will fast-track vehicle development. In the absence of having autonomous vehicles in Europe tomorrow or shortly thereafter, we will have automatic, communicating, extremely reliable and sophisticated cars, even at entry level.
With the growing automation of cars, do you think that the pleasure of driving, and the sense of freedom that goes with it, will always exist?
This feeling of freedom has already virtually disappeared: everything possible is being done to curb it. Over time, we have become safer, which is a positive thing, of course. Systems are being developed for us which will automatically prevent us from exceeding the speed limit. You will always be able to deactivate them, but they will be activated by default. If you decide to deactivate them, it will be noticed, and it will count against you in the event of an accident.
When I started in this business, 30 years ago, driving was pleasurable. There weren’t too many speed cameras, you could drive fast in certain places. Nowadays, the moment you exceed the limit by as much as 1 km/h, you are a potential offender! Everything is done to restrain you, to compel you to comply with every rule to the letter and the trend is not going to reverse.
Ultimately, the only place you will be able to rediscover this feeling of freedom and have a bit of fun will be in video games and driving simulators!