The road to an autonomous future: within reach, or a tantalizing mirage?

Although a fully autonomous vehicle fleet seems closer than ever, the remaining hurdles are daunting, writes Kevin Roberts

The rapid rate of innovation on technologies such as artificial intelligence has shifted the narrative regarding autonomous vehicles (AVs) from whether such vehicles are even possible, to how fast companies can bring commercially viable AVs to market. Complicating matters is COVID-19 and the impact on how people view mobility going forward; there are now serious questions regarding the long-term prospects of mass transit and shared mobility. The question now has become: can the AV be a solution to these challenges or will it be a casualty before it even launches?

Special report: What’s the timeline for autonomous vehicles?

The case for AVs

The biggest case for an AV is safety. While human drivers have been improving with regard to safety over the intervening decades, helped largely by technology advancements, there remain far too many traffic fatalities. The World Health Organization (WHO) estimates that in 2016 there were 1.35 million road traffic deaths globally; that averages out to around 3,700 people killed daily. The US Department of Transportation estimates that 94% of all serious motor vehicle crashes involved a driver-related factor, with the remainder being caused by environmental or technical issues. We could effectively reduce that number to near zero with AVs.

EY-chart-1 - road traffic fatalities

Beyond the potential to save lives, there are other benefits to AVs, such as the amount of time that is lost sitting in traffic. According to the 2019 Urban Mobility Report, the average commuter in the US spends 54 hours in their car in congestion; that time could effectively be freed up with the advent of AVs, allowing for more work or leisure.

The reasoning for governments and consumers seems clear, but what’s in it for the companies developing the technology? How about US$8tr? Recent industry estimates put the global AV industry as an US$8tr opportunity, with ride-hailing potentially accounting for US$5tr, US$2tr for freight, and US$500bn each for data insights and in-vehicle experiences. That US$8tr is a significant increase over the automotive industry today which, according to the International Organization of Motor Vehicle Manufacturers (OICA), accounts for roughly US$2tr in gross revenue.

What’s the hold up?

If the above has you interested in an AVs, you’re likely to be wondering when the technology will be available on the road for you to use. The answer—unfortunately—is not an easy one. What one means by autonomous—how to connect to these vehicles, use applications, regulatory uncertainty, and consumer concerns—currently cloud the outlook for when a commercially viable AV will be on the market. While the causes for concern are many, by looking at each in detail, we can start to get a better understanding of the timeline for AVs.


The road to full automation

The move to AVs is unlikely to be a binary one as we see more and more levels of autonomy added into our vehicles today. In fact, vehicles today already feature many advanced driver assistance systems (ADAS) which are helping to make driving safer today while decreasing the need for and risk of human drivers. We are already seeing technologies such as collision warning jump from being installed on just 13% of 2015 model year vehicles in the US to 60% in 2019 model year.

However, this evolution in safety is also creating issues for both drivers and vehicles. Solutions for Level 3—conditional automation—create potential issues regarding how long a driver can hand off operation for and when they need to take back control. These gaps in human driving versus automation can create situations where a driver becomes disengaged from the activity and could increase instead of decreasing risk.

Instead, we’re starting to see an emergence of Level 2+ functionality—partial automation—which increases safety but does not allow the driver to remove themselves from operational completely. The question now is whether we can move from Level 2+ to Level 4—high automation—directly.


Connected car

While there remains uncertainty around the timing of a safe leap to Level 4, there is no doubt that we’ll need to have reliable, pervasive, high-speed connectivity for this to take place. The need for such connectivity is critical to underpin vehicle-to-everything (V2X) communication, which would allow AVs to take advantage of numerous connection points to provide improved levels of safety to all, improvement in entertainment for riders, and more efficient travel reducing congestion and emissions, helping to not only reduce environmental impact, but boost the economy. By one estimate, congested roads in the US cost the economy US$871bn annually, a cost which V2X could help dismantle.

The right technology to allow all these pipes to interact will be critical. There is hope that 5G could be the technology to allow this to happen; however, with its roll-out just beginning, we cannot expect full global use of 5G immediately. Additionally, while 5G is an improvement over the current 4G system, most people would not want to chance riding in a fully autonomous vehicle with the connectivity issues that cellular connections are known for, so further redundancies to ensure no connectivity losses will be critical.

EY-chart-2 - V2X technologies

The right AV for the job

We tend to view AVs as a monolith of sorts; however, there are several potential options that are being developed and some are likely be brought to market sooner than others:

  • Mobility-as-a-Service (MaaS) – When most people think of AVs, MaaS is likely to be what springs to mind: a driverless ride-hailing vehicle that can be summoned from a phone and take them where they need to go without the inconvenience of having to drive. This is also the most complicated version of AVs to come to market, and faces a litany of technological, regulatory, and customer concerns before they can successfully come online. Additionally, the technology is likely to be limited to select markets in heavily geo-fenced boundaries.
  • Commercial vehicle – While MaaS might be outside the mainstream, we are likely much closer to Level 4 autonomy on freeways for commercial vehicles. This space would offer a quick return on investment, with the driver taken out of the equation for the long-haul component and then re-inserted for non-freeway driving.
  • Last mile delivery – Recent approvals for technologies in certain markets such as Nuro, allowing the testing of its autonomous delivery vehicles on the streets of California, or Starship autonomous deliveries in Milton Keynes in the UK mean that soon an AV could be delivering groceries, replacing curbside pickup.

Concerns remain

Beyond technological concerns, there remain many questions regarding how AVs will be regulated and who will do so. If each individual country regulates independently, it will likely slow down global roll-outs, as software will have to be customized and tested for each jurisdiction. This is a thorny issue, with seemingly few easy answers, but it is crucial for the cause of AVs that it is addressed before too long.

Finally, there remains a question regarding whether consumers want an AV at all. A recent study by AAA found that 71% of US drivers would be afraid to ride in a fully self-driving vehicle. So, beyond technological and regulatory hurdles, there will also need to be a campaign from manufacturers, automakers, tech companies and governments to convince consumers that an AV is safe for them and their loved ones to use.

The enormous potential and possibilities that a fully autonomous vehicle fleet would bring to the market seems to be tantalizingly close, yet the remaining hurdles that must be overcome to make it a reality are daunting. The additional challenge of COVID-19 clouds the outlook for AVs even further, as a shift toward more private ownership could dim the future of autonomous shared vehicles. Can all of these trials be overcome in this decade? Perhaps, but not all at once. What will likely occur will be an accretive path to an autonomous future that will be highly application- and region-dependent.

About the author: Kevin Roberts is Associate Director – Automotive & Transportation Senior Analyst at EY