AV Global Readiness Index: Who Will Get There First?
The AVRI and Why It Matters
The Autonomous Vehicles Readiness Index (AVRI) is a score devised by KPMG which assesses a country’s current readiness for incorporating AVs onto their roadways. The AVRI rates these countries on a number of different axes, from innovation to infrastructure.
In the current version of the index, the United States ranks fourth behind Singapore, the Netherlands, and Norway. However, it ranks second in innovation behind only Israel. This makes sense given the number of startups and established companies that the US has working on autonomous driving.
In fact, the US is home to 420 AV company headquarters including the likes of Tesla, Cruise, Waymo, Drive.ai, Argo AI, and many others.
AVRI Current Top-10 Rankings
- The Netherlands
- United States
- South Korea
- United Arab Emirates
- United Kingdom
US AV Infrastructure and the New Infrastructure Bill
In contrast, the US performs somewhat weakly on infrastructure, especially when compared to European countries. The Infrastructure Report Card is a great resource for learning more about the current state of America’s public transit, roadways, and other public works systems. In particular, the US scores a “C-” overall according to the report card’s rating system, and a “D” for roadways.
This is likely largely due to the fact that, according to the report, 40% of America’s roads are in poor-to-mediocre condition. One reason for the US’s crumbling infrastructure is that the US taxes less than many other countries, particularly those in Europe. Another difficulty in the US is that the infrastructure quality varies from city to city depending on how supportive the local and state governments are of AV efforts and tax dollar reinvestment, among other things.
Takeaway #1: Public Transportation
One area of infrastructure which stands to be significantly impacted by the new bill is public transportation. Outside of a few major cities such as New York and San Francisco, most of the US lacks access to robust, affordable public transit. AVs could certainly change that for the better.
About $39 billion of the bill is earmarked for public transit improvements, with the majority of that money intended to go towards new projects. AVs act as a complementary force to other systems in current use such as subways and metro buses. While the $39 billion in funds have not yet been explicitly assigned, perhaps a portion of it will be used to invest in self-driving buses or autonomous taxi services which could reduce the burden of cars on the road.
Takeaway #2: Roadways
As noted in AI on the Fly, the bill also has a strong focus on improving the quality of US roadways. As noted by the AP and the New York Times, $110 billion of the trillion dollars is specifically devoted to revitalizing aging roadways, bridges, and other key infrastructure. This could potentially fast-track the US’ readiness for AVs by smoothing out obstacles in the driving environment.
For example, with fewer potholes, fading lane markers, and other hazards on roadways, the self-driving “brain” of a car will have fewer distractions to contend with and will be able to focus more on other important objects in its environments, such as pedestrians. However, it will be interesting to see if improving the quality of roadways might actually lengthen somewhat the time until AVs are ubiquitous on our streets.
Since current AVs have been trained entirely on images and video of our existing roadways, deploying them in environments with significantly altered infrastructure might require gathering new data and training the model. We are not aware of any study that has investigated the potential impacts of such decisions, so it will be fascinating to see how it all plays out.
Takeaway #3: 5G and Connectivity
The bill also stands to devote a portion of its total funds ($65 billion) to improving internet connectivity within US cities. This includes building many more 5G towers, expanding fiber internet lines, and much more. With increased connectivity, AVs will increasingly be able to interact with the world and each other, bringing about a number of improvements.
For one, AVs will be able to continuously stream data and analytics from each of their drives to the cloud. This will allow them to alert other nearby vehicles of road hazards and traffic conditions and will improve the ability of a central scheduling system to coordinate the flow of vehicles through potential congestion points. AVs might also gain the ability to retrain their internal models on the fly, thus allowing them to gather information from the world and learn from it, all in real-time.
Takeaway #4: Cybersecurity
However, increased connectivity also brings with it some risks. Any time a computer opens an internet connection to the outside world, it becomes vulnerable to attack, and AVs are no exception.
There is concern that as the US expands the connectivity of its vehicles, rogue actors will look to exploit the weaknesses in AI systems, perhaps by feeding them adversarial examples that will cause them to break traffic laws (e.g. blow through red lights or stop signs), crash, or even attack pedestrians. Therefore it is vital that the US also invest heavily into security implementation and research as connectivity ramps up.
Doing so is the only way to responsibly ensure that our roadways remain safe and free from external influence.
Takeaway #5: EVs and Clean Energy
In addition to these improvements related directly to self-driving tech, the bill also stands to impact AV readiness in more tangential ways. For example, it allocates $7.5 billion specifically for the building of electric vehicle charging stations and another $65 billion to modernizing the electric grid as well as investing in the development of new clean-energy technologies such as hydrogen fuel cells.
This will all be vital for getting AVs onto our roads as these autonomous systems are being developed for the next generation of vehicles, those that are likely to run primarily on hydrogen and electric. Indeed, Tesla is more or less leading the charge in terms of having actually self-driving tech deployed in the real world, and one major consumer deterrent to purchasing Tesla vehicles is the “range anxiety”, i.e. the anxiety associated not knowing if you’ll be able to find a way to charge your vehicle when driving from point a to point b.
Therefore, building a much more robust charging infrastructure will be vital to getting autonomous vehicles out on our roads.
The $1TN infrastructure bill carries many implications for modernizing the infrastructure that will ultimately result in the widespread adoption of autonomous vehicles and mobility. While much of this is undoubtedly useful for the adoption of autonomous vehicles, manufacturers and innovators will still need to master the art of crafting these models for them to perform on any road, modernized or not. While infrastructure is a key element to the success of autonomous vehicles, it’s only part of the equation, with high quality data and masterfully executed AI models comprising the other part.
For more information of scaling data pipelines, conquering edge cases, and creating autonomous models that can perform on any road, download the iMerit solutions brief today.