Are artificial cities the answer to AV road safety?
Technology and Business
In this article, Alice explores artificial test cities and asks:
- What are the advantages of artificial test cities?
- Are these set-ups a better option than virtual simulation for testing AVs?
- What do the experts have to say?
Roads are unpredictable places. Though there are plenty of rules in place for conduct on every highway in every part of the world, we can’t rely on drivers to follow them to the letter. And it’s not just drivers that can cause confusion. Battered signs, worn road markings and unexpected obstacles all pose a challenge.
So, it goes without saying that driverless vehicles need to be prepared for every eventuality and we know that safety is an important factor to be addressed when it comes to AVs. As a result, rigorous testing and real-world (or replicated real-world) experience proves vital in swaying public opinion of, and trust in, the tech. While you could drive for days without encountering some of even the most common obstacles you’ll find on the road, in controlled environments like artificial test cities, technicians can guarantee them. This proves vital in forming public opinion of, and trust in, the tech. While you could drive for days without encountering some of even the most common obstacles you’ll find on the road, in controlled environments like artificial test cities, technicians can guarantee them.
What are the advantages of artificial test cities?
As Stephane Barbier, chief development of officer at Transpolis, a mobility testing ground at the foot of the French Alps, explains, “To statistically prove an autonomous vehicle is as good as a human driver, it would have to be road-tested for millions of kilometres. Fake cities, complemented by virtual testing, are the best way to rack up the testing kilometres without putting lives at stake.” Several campus-style test cities which replicate real-world world routes and common obstacles now exist across the globe.
South Korea’s K-City was among the first and is now integrating communications into the testing it conducts with the implementation of 5G. Mcity at University of Michigan, though, is perhaps the best-known artificial city. Covering 32 acres, it features everything an AV might encounter on the road including worn road signs, ‘pedestrian’ dummies and even a tunnel which mimics overpasses where signals may be disrupted.
But before reaching one of these artificial cities, most AVs will first undergo other forms of assessment. Simulation is a common step which falls before testing on an artificial city track and pilot programmes in live testbeds usually follow. So why are artificial cities needed in between?
Are these set-ups a better option than virtual simulation for testing AVs?
Simulation is used at some point in the development of all driverless vehicles and massive leaps are being made in terms of what the process can offer. Tech company, Nividia, for example, have developed NVIDIA DRIVE software and simulators to offer a scalable, comprehensive, customisable and diverse testing environment without the need for the travel, expense and man hours associated with live testing.
Software specialist rFpro have also developed their offering recently, and hope that improved accuracy will mean simulation could become a significant part of the approval process for AVs. Matt Daley, rFpro’s managing director, explains, “You cannot feasibly validate autonomous vehicles in the real world alone; there simply isn’t enough time or money. We must also exploit the expandability of simulation.”
You want to know more about how the recent Covid19 has affected the journey of autonomous technology on our roads? Read our article "COVID19: smart mobility goes viral."
It’s clear simulation has its place, and brings huge benefits and opportunities with it, but in terms of developing public trust, it simply can’t match testing in a real physical environment. Plus, now that much artificial city testing is blended with simulation to some degree, this type of on-road assessment offers more for engineers.
What do the experts have to say?
With that in mind, we spoke to Russell Vickers, CEO of Future Mobility Campus Ireland (FMCI), a newly opened artificial test city. Designed as a collaborative testbed which offers manufacturers the opportunity to trial new models on 12km of roads simulating all sorts of environments, traffic scenarios and potential obstacles, according to Russell, FMCI is the latest, and arguably most advanced, test city we’ve seen to date.
Why are testbeds like FMCI so important for the development of autonomous tech?
As you can imagine, there are some really difficult and expensive questions to answer when it comes to the future of mobility. This is not just related to autonomous driving, but how to manage the data, connectivity and services necessary as we build out transportation systems for the future. Because of this, no one company will be able create these solutions in isolation. The Future Mobility Campus Ireland (FMCI) is a platform that will allow entities to leverage the testbed facility to create new products and companies without the prohibitive costs to procure and install the necessary infrastructure to build out these new innovations.
Do you see artificial test cities as a one-stop solution, or part of a broader testing process?
The development of these technologies, particularly automated driving is always a staged process to ensure safety. Beginning with virtual validation, then moving to private track testing followed by real-world testing on public roads with strict safety conditions until the system is fully validated. So all pieces of this puzzle are required, 95-99% of the feature under test can be validated without public road testing but you cannot get to 100% with this key aspect. Aside for the automated aspect though, having a real world setting is very important to developing any type of product/solution that will form a part of future mobility.
Why is real world testing important? Only for technological reasons?
There will always be use-cases that you cannot simulate and these need to be proven in the real world. For sure, simulation is a growing part of validation for these technologies, but you really need to run these products in the wild to understand how they will interact with their environment and, more importantly, how the environment will react to these technologies. For me, the sociological aspects of how we as a society interact with this potential change is really interesting.
What would you most likely do with your free time in an autonomous car?
At this moment in time, I would probably need to answer all my emails related to FMCI! After that, probably catch-up binge watching on Netflix, and maybe even enjoy the view out of the window.
Collaboration, it seems, will be a major theme for the development and testing of AVs as automotive manufacturers, tech companies and testing facilities must work together to ensure vehicles are safe for passengers and fit for the future. Coopetition, as co-operation between competing companies is sometimes called, will be a major theme for the future of driverless. Though artificial cities designed just to test AVs may sound like a big investment, and can be ambitious projects in themselves, it’s clear they’re having a huge impact on progress in the field of driverless tech. We’re eager to see where they lead us.
Your opinion matters to us. What do you think of artificial city testing? Do you think they could ever replicate real-world scenarios? Do you think they could replace real-world testing? Share your thoughts in the comments below.
COVID19: smart mobility goes viral
Will everyone win in a driverless future? - An expert’s perspective
Submit your story
Become part of our autonomous revolution and submit your stories, images and videos
Stay up to speed with our weekly briefing. Enjoy autonomous driving content direct to your inbox