In April 2016, Ford and Volvo joined forces with Google, Lyft and Uber to form the Self-Driving Coalition for Safer Streets. The latter three may be comparatively new players in the automotive industry, but they are deemed to possess the skills required to make self-driving vehicles a reality.
Fronted by David Strickland, the former Administrator of the US National Highway Traffic Safety Administration (NHTSA), the Coalition is under the direction of a seasoned vehicle and road safety advocate. Specifically, he will serve as the Coalition’s Counsel and spokesperson, responding on a moment-by-moment basis to the challenges of deploying autonomous vehicles in the US and beyond.
“The long-term goal is to have a regulatory environment to allow the full deployment of self-driving vehicles where there is no expectation for the driver to intervene in the driving task,” he tells Megatrends. For Strickland, there is no time to lose.
For the first time in a very long time, road fatalities in the US have increased year-on-year. According to NHTSA, the number of road deaths in 2015 rose 7.7% to 35,200, compared with 32,675 in 2014; NHTSA data also shows that 94% of all road collisions are a result of human error.
“We have been fighting this war against crashes and fatalities for decades, and we have been generally successful,” he says. “However, our reductions have not been as significant as those in developed countries in Europe and Asia.” The US has work to do in Strickland’s eyes, and the Coalition’s efforts could prove instrumental in encouraging change.
The expectation that a self-driving car will never make an error, ever, is ludicrous
Automated driving technology is not only expected to drastically improve road safety by eliminating driver distraction and poor decision making, but also provide those with limited mobility the access to on-demand transportation. As Strickland puts its, driverless cars are “the new Holy Grail” for the automotive industry.
Blanket regulation
First on the list of duties is to address the regulatory framework surrounding automated vehicle technology; namely how, where and when it can be tested and eventually released on public roads.
The Coalition – along with most stakeholders – is pushing for a national regulation that carries across borders to allow the free-flow of autonomous vehicles from one state to another. A patchwork of regulations would make smooth national deployment of autonomous vehicles difficult; a single blanket regulation cut from the same cloth is the objective, he explains.
“The goal is to try to get it right at the first cut so that we can unleash innovation in the most efficient way, but always keep our eye on the fact that uniformity doesn’t mean a weakened or a lesser standard,” he says. “It shouldn’t be the Wild Wild West out there. There should be some notion of uniformity to make sure that everybody – once the rules are established – has an even playing field on which to deploy and compete.”
He observes that there may eventually be small differences as certain states have varying police, traffic safety and enforcement laws. In addition, not every locality in America allows cars to turn right on a red light, for example. “There’s going to have to be some recognition of differences at certain levels,” says Strickland, and suggests that it would be simpler to tweak a group of variations than to have firm limitations depending on where a driver is in America.
“If you have a self-driving vehicle that is licensed for testing in California and want to bring the vehicle over to Nevada, you drive it to the California line, change over to Nevada plates and then go across. That’s the only way to be legal,” he explains. “If that’s how individual states treat self-driving vehicles, it’s going to be really hard to be able to get the full benefit.”
Five cohorts
The five partners of the Coalition are all firmly involved in autonomous driving in some form.
Ford has been carrying out extensive testing at Mcity, a mock-up town in Michigan used to replicate various driving situations. The OEM has also been testing its self-driving technology in winter conditions and in absolute darkness.
Google began the development of its self-driving technology in 2009, and has since racked up just under 1.8 million miles of autonomous driving on public roads. Ride share platform Lyft recently partnered with GM with the aim of deploying autonomous fleets in the next couple of years.
Volvo is gearing up to launch the public aspect of its Drive Me initiative; 100 vehicles equipped with self-driving technology will be driven by members of the public in Gothenburg, Sweden in 2017. In May, Uber announced it would launch a hybrid Ford Fusion test car in Pittsburgh to test self-driving technology and gather mapping data, and in August it unveiled plans to work with Volvo on autonomous vehicles, including a fully autonomous car that Uber intends to use in its alternative taxi fleet. That announcement came shortly after Ford committed to launching a high-volume, fully autonomous SAE level 4-capable vehicle by 2021.
Ride-sharing, or private ownership?
While the inclusion of two ride share companies would suggest the Coalition is pushing for autonomous taxis, Strickland explains that this is not the only application with a promising future.
“Ride-sharing is one application that has tremendous promise, but Ford and Volvo Cars are also members of the Coalition, and they have plans for private cars with full self-driving ability,” explains Strickland.
However, he suggests that autonomous ride-sharing will likely arrive first in meaningful volumes. “[Private ownership] will be further out in the future than a ride share application in a congested megacity.” He suggested that, similar to Google’s test cars, there will be a fleet of self-driving vehicles that can go no faster than the internal speed limit of 25mph (40kph) in a downtown area: “That’s probably going to be one of the earliest applications for people to be exposed to self-driving cars in a consumer environment, and then it will grow from there.”
In the long-term, Strickland believes automation will be available through a number of applications, from ride-sharing to private vehicles that can drive independently in all areas and weather conditions. The latter is “ultimately what folks are aiming for,” he muses.
A fatal incident
On 7 May 2016, the first death linked to automated vehicle technology occurred in Florida when a Tesla Model S in Autopilot mode collided with an 18-wheel truck. It was reported that the semi-autonomous system could not distinguish between the white paint of the truck and a bright skyline, and as such did not slow down or try to avoid the collision. While safety advocates have called for the system to be renamed and deactivated until deemed ‘safe’, others believe this was an incident waiting to happen.
It is worth noting that this was not the first time autonomous vehicle technology has been involved in a collision on public roads. An autonomous Google car was involved in a low-speed collision with a bus in February 2016, and according to the US Department of Motor Vehicles (DMV), a Nissan Leaf fitted with Cruise Automation technology had crashed into a parked car in San Francisco a month earlier.
We should not let a crash due to a self-driving vehicle or ADAS feature pause or delay our movement towards autonomy. That technology really is the most direct answer to reducing human error behind the wheel
Speaking at the Autonomous Car Detroit conference in March 2016, Strickland’s successor and current NHTSA Administrator, Dr Mark Rosekind, spoke of not letting the pursuit of ‘perfect’ autonomous driving technology get in the way of deploying ‘good’ technology that can still save lives. However, in the wake of the Autopilot incident, Rosekind received a letter from four safety advocates, including another former NHTSA Administrator, Joan Claybrook, that called this ideology a “false dichotomy”.
“No technology can deliver on a promise of safety if it is rushed into vehicles with known deadly defects,” the letter reads. However, Strickland is unconvinced that the event in Florida will change how regulators look at the deployment of autonomous drive technology.
“Administrator Rosekind has already spoken to the fact that while the Tesla crash is immensely tragic – much like every crash that leads to a fatality – 35,000 people lost their lives in crashes in 2015. The Model S crash was, unfortunately, one of way too many in 2016,” he says. “As a former Administrator myself, my perspective is that there was always going to be a situation where an advanced driver assistance system (ADAS) or self-driving system was involved in a crash that led to a fatality.”
He adds that although it was not a question of if, but when such an incident would occur, “it probably came earlier than we all would have thought.”
Sooner rather than later
Ultimately, Strickland agrees that given the rising number of road deaths in the US and the fact that nearly all can be attributed to human error, the deployment of autonomous driving technology should continue without delay.
“We should not let a crash due to a self-driving vehicle or ADAS feature pause or delay our movement towards autonomy. That [technology] really is the most direct answer to reducing human error behind the wheel,” he affirms, but notes discomfort with the concept of ‘perfect’ autonomy. “No vehicle system ever created is 100% safe. You can build countermeasures, but in the history of automotive safety we have seen that in the multivariable notion of chaos, unintended things happen and you have to address that,” he explains. “The expectation that a self-driving car will never make an error, ever, is ludicrous.”
Administrator Rosekind has already spoken to the fact that while the Tesla crash is immensely tragic – much like every crash that leads to a fatality – 35,000 people lost their lives in crashes in 2015. The Model S crash was, unfortunately, one of way too many in 2016
That is not to say that ‘near perfect’ is unattainable. Six Sigma is a data-driven methodology that aims to eliminate defects in any given process. Strickland expects that through this approach, the industry can strive to develop autonomous vehicle systems that are 99.99% effective. “You want to make sure of that, just like any good automotive manufacturer would do with any safety system… But you can’t tell consumers that anything is 100% safe because it is just not true,” he admits.
Sitting and waiting for ten to 15 years for a bullet-proof system is foolhardy, he concludes. “Doing so would leave benefits at the table. We can help keep people safe faster with this technology being deployed.”
This article appeared in the Q3 2016 issue of Automotive Megatrends Magazine. Follow this link to download the full issue.
