What happened to Autonomous Vehicles Acceptance criteria?

6 min read

Looking back at 2013, when AVAIRX was supporting the development of two of the now famous autonomous shuttle manufacturers in Europe, what ever happened to autonomous vehicles predictions for operations planned across streets of the world by 2021, whether for carsharing, for private use, for ride-hailing or replacing public transport and on-demand transport?

Perspectives of autonomous vehicles were very optimistic!

They were led by car OEMs preparing to fight back tech giants already proposing driverless cars for years, such as the 2015 Google Car , now, Waymo , part of Alphabet in Googles’ constellation of innovations.

Disrupting use-cases were also wild, sleeping on the highway for long distances, competing with regional aviation and train, reading, working, and attending video conferences, much before the Zoom -era.

Concern was already to categorize autonomous vehicles, guided but how, and providing what security compared to existing guided vehicles.

Ever since humans mastered fire, we have been relying on the idea of being in control, from guiding a horse-driven carriage, to the joys of driving your own car everyday.

The AV technology was not the problem, OEMs pointed to onboard sensors, road infrastructure and intelligent transport system-providers pointed to Vehicle to Infrastructure V2I, Vehicle to Everything V2X and the digitalization of transport.

Then came the tsunami of Field Operational Tests across the world and countries, territories even competed to attract all kinds of tests, depending on weather conditions, topological variations, the potential wildlife commonly crossing roads (Moose in Scandinavia, kangaroo in Australia, etc), differences in the urban fabric, integration with public transport, in airports, ports - it all summarized to who would develop the best Mobility Valley to welcome autonomous vehicles trials.

A question remained - “How to deploy AVs safely?”

In the acceleration of emerging and non-conventional technology, there always comes a time when users and clients doubt, putting the general acceptance at test, from perceived usefulness to perceived ease of use, towards system usage and the long sought-after realization, referred to as “Jitsugen” in the Japanese society- the proper, actual, effective roll-out and usage of innovation.

実現 ("jitsugen" meaning "realization" in Japanese)        

This realization was challenged with the first casualties. Human fatalities came to the spotlight in March 2018 with Elaine Herzberg , the first recorded case of a pedestrian fatality involving a safety driver, the driving environment, software and sensor issues and human distraction, as the safety driver relied on the vehicle’s “intelligence to provide early collision avoidance warning and the ability to regain human control over the system”.

As this project was developed by UBER , one of the major and the most famous ride-hailing operators, the ecosystem took a blow and FOTs started to become more complex to organize. Governments started requiring more safety guarantees. In December 2020, UBER’s Advanced Technologies Group (ATG) was acquired by Aurora through a strategic partnership.

This trend led to an autonomous vehicle ecosystem focused on safety and security, partially putting aside business and operation related criteria of AVs, while focusing on limited public acceptance R&D, with less than 10 projects actually covering the vast array of AV acceptance criteria, studying what business could arise from a full AV deployment in transport systems.

Although UBER and Lyft gave up their AV projects, 2020 showed more positive signs as TESLA rolled out more semi-autonomous Autopilot functions. The latest automated driving technology was made available for delivery robots during the COVID-19 first lockdowns, especially in China, the USA and autonomous shuttles won major subsidized projects in Europe following their autonomous-demonstration strategy everywhere.

The autonomous market got investors and the public nervous again, and the ecosystem wants AVs to come up to speed again. New kids-on-the-block come in and become stronger such as Zoox , purchased by Amazon , quoting the future "new normal"; May Mobility , backed by Toyota thanks to early revenues of approximately $4/ride/passenger.

For the near future, clients of AVs will still have to keep their hands on the wheel and their eyes on the road even while changing lanes.

Some major changes since 2013

The curves are autonomously and gently driven, with less emergency braking, coding mistakes are rapidly identified and corrected, AV road-driving runs more comfortable, and the general public has now the legal right to drive level 3 vehicles in a number of countries. Elon Musk is more and more confident that Tesla will soon be able to offer a fully autonomous car despite describing it as one of the toughest challenges ever.

The promise of creating hyper affordable Robo taxi services and getting rid of 60% of the operation cost of a driver was an oversell. It also created unreachable expectations, as stakeholders and regulators were unaware of limits of use.

The AV business is now dominated by companies such as Toyota , GM and Ford , other tech companies and car manufacturers like Cruise a subsidiary of General Motors have not given up and are pouring billions of dollars developing autonomous vehicle FOTs for robot taxi services in China, Russia and the USA.

Despite the hype and today’s soft landing, the last years have seen progress in automated driving technology. New partnerships arise such the one between Ford and Argo AI , cooperating with the Lyft ride-hailing network.

The objectives have shifted from self-driving cars to automated and autonomous vehicles, with a clear acknowledgment of the name “autonomous vehicles” that can be deployed in cities, complex rural and suburban environments.

Experts and stakeholders’ rate that another decade will be necessary to improve machine learning, sensors fusion and computing power for all these trends to reach commercialization.

Safety remains critical as our surveys have shown, acceptance is kind of the abandoned child of the family. Early tests like the DARPA Challenge helped learn and get more engineers to engage.

The last decade helped make the general public aware of what’s coming, the next decade will be to get them on-board to validate a number of acceptance criteria.

What acceptance criteria to choose and develop?

It is not anymore about machine learning, object recognition and detection, sensoring a bicyclist and avoiding it, not being stopped by an uncategorized dead leaf. It is now about predicting what acceptance criteria the client will be focusing upon beyond safety, as safety and security will be a sine qua non.

Should it be about the brand offering an AV, the trust in the system designer, as much as people trust Apple when Foxconn assembles devices. Will customers accept the same with their new MIH-EV platform, the willingness to pay a service, owning an AV car or not, a specific individualized AV service for dedicated use cases under new city regulations... Will customers trust Apple’s autonomous vehicles when they partner with an OEM like Hyundai for instance, while Foxconn builds its story telling on harmony and an open ecosystem as a criteria?

Skepticism about the technology and time needed for providing services to clients means there is still a lot to be confident over the 2 next decades. Experiencing AVs in all kinds of environments and situations becomes certain with limited set of capabilities and considering limits of use.

Human evolution has gone through similar upheavals, this challenge is greater as this means humans will accept to yield and give control to machines, developing specific regulations to provide exemption of liability to humans.

The trends in the mobility ecosystem are undergoing major tectonic shifts, and the acceptance of automation as such remains a gigantic challenge for autonomous cars. For them to be widely accepted will also mean finding an alternate source of income for people currently depending on driving for a living.

The challenge will not be about millions of test miles driven and crashes avoided, but, following a centennial crisis, what jobs and livability will AVs provide to communities and the transport ecosystem.

As for acceptance, playing on new criteria, building on general public’s acceptance criteria while reminding of conventional safe rides, some companies, such as Ford , have come to extremes in their marketing on how boring an autonomous hands-free ride maybe, while driving on a highway, heading for holidays.

Will societies need to provide a job loss insurance whenever a driving position is removed, rather than a car driver’s insurance?

What invariants and parameters will humans consider adopting the technology?

How will humans react to the expected superhuman capabilities of AVs??

What benefits to society will AVs bring?

Will electrified AVs contribute towards net zero transport for both people and goods?

Will they create new jobs?

What insurance company will be ready first to offer a premium, not linked to a car safety, but to a Service Level Agreement, like arriving on time at destination thanks to an AV?

How much will humans be willing to accept handing over control to AVs as much as we gave the control of fire to nuclear power plants? Or will we see workers destroying AVs factories, as tailors, back in the 1830s, worried about being put out of work, destroyed sewing machines workshops.

Find out more in the AVAIRX Autonomous Vehicles Acceptance State of Play 2021 and find out who is on track and at the Avantgarde of the next criteria to boost AV acceptance in the World.