Years before Ray Kurzweil became Google’s director of engineering, he invented a line of musical synthesizers, the first flatbed scanner, the first optical character recognition, and the first commercially viable large-vocabulary speech recognition—among numerous other achievements. The Wall Street Journal calls him “the restless genius” and Forbes Inc. “the ultimate thinking machine” and “the rightful heir to Thomas Edison.” In his recent book, entitled The Singularity is Near, Kurzweil describes a universe made not of energy nor of matter, but one made solely of data.
In the automotive world, the notion of Kurzweil’s universe of data can’t be avoided. Transportation is becoming a universe of data as vehicles themselves gain computing power at a fantastic pace. According the The Boston Consulting Group we have reached an inflection point where safety systems are morphing into fully autonomous vehicles. A journey that begins with today’s collision avoidance systems is heading fast toward the horizon of self-driving cars.
Volvo, for example, has publicly vowed that within five years not a single person will be killed or injured in a Volvo vehicle. They make that promise with confidence backed up by a broadening spectrum of increasingly effective sensors and related technology.
Volvo’s City Safety, a low-speed frontal crash prevention system installed on 2011-2012 S60 and XC60 has been credited with a 15 percent reduction in claims under property damage liability in the US, according to Highway Loss Data Institute reporting. Those results are not atypical when it comes to other frontal collision avoidance systems.
In separate analyses of front crash prevention systems that function at higher speeds, HLDI found 14 percent fewer insurance claims under property damage liability coverage for Acura and Mercedes-Benz vehicles with forward collision warning with automatic braking than for the same vehicles that weren’t equipped with the technology.
HLDI also looked at adaptive headlights offered by Mazda, Mercedes and Volvo and found property damage liability claims fell as much as 10 percent.
Yet another study of large trucks concluded that blind spot detection, forward collision warning, lane departure warning and electronic stability control together could prevent or mitigate as many as 28 percent of large truck crashes, including one in five fatal accidents.
Collision avoidance systems are simply becoming smarter and relatively less expensive, and expected to continue to gain fleet penetration. Global sales of anti-crash sensors will total $9.90 billion in 2020—up from $3.94 billion last year, predicts IHS Automotive, a research firm based in suburban Detroit. Bosch sold about one million radar units in 2011. In 2015 and 2016, the German supplier expects its annual sales will average four million units.
In Ray Kurzweil’s mind, people tend to under-estimate rates of change. Change occurs, he notes, not as an often expected linear trend but as a series of “S” curves on a graph. As a trend gets established, it accelerates for a time before flattening out. When people assume a constant rate of change, they are often taken off guard by the acceleration segment of a change curve.
Intelligent vehicles systems have arguably now reached the beginning of the steep part of the “S” curve, as Volvo’s recent promise to eliminate fatalities and injury by 2020 illustrates. That’s little more than four years away.
At least one forecast has fleet penetration for frontal collision avoidance systems at a lower limit of 40% within seven years. A key result of this substantial reduction in accidents—a collision repair market 15% smaller than today.
Rapid change on an “S” curve is often characterized by little leaps in technology. Delphi Automotive, for example, has announced it will produce radar for two automakers introducing 360-degree vehicle surveillance. It would seem likely that we have yet to see the last of such innovations.
Certainly, the advance of self-driving cars will run into its share of roadblocks, many on the regulatory side. Yet few doubt that we are within a generation of fully autonomous transportation.
As we move in that direction, the next few years will see rapid market penetration of systems that can take more and more control away from drivers. Lane departure and adaptive cruise control available on new models provide provide assistance in stop and go traffic by maintaining a safe distance between vehicles and adjusting speed through both braking and acceleration to better handle on crowded roadways.
Volvo, for their part, plans to introduce intersection braking and rollover protection on passenger vehicles this year, and further plans to have the first fleet of 100 semi-autonomous passenger vehicles on the road in Gothenburg in 2017.
For 2017 Cadillac will offer “Super Cruise” technology that takes control of steering, acceleration and braking at highway speeds of 70 miles per hour in stop-and-go congested traffic.
Mercedes already sells a system that drives a car on the freeway at speeds of up to 120 miles per hour, as long as the driver keeps a hand on the wheel.
BMW’s currently available Traffic Jam Assistant controls the speed of the car and distance to the car ahead in dense traffic on motorways at speeds of up to 60 km/h, and even takes over steering.
Tesla Motors Inc. plans to release a car with “auto- steering” this summer. Google will see a number of its self-driving cars leave the test track environment and hit the road at the same time. The new prototypes operate on the same software that Google’s existing fleet of Lexus RX450h SUVs uses. That fleet has already logged nearly a million autonomous miles, and have been covering 10,000 miles per week.
Lane-keeping and automatic braking systems can now be found on mid-market models from Honda, Ford, Hyundai. and others. Driver-assist packages that a few years ago cost owners $4,000 or more, now can be purchased for less than half that. Lexus has said it will offer a suite of semi-autonomous safety features on the next RX SUV for under $650.
By 2017, partially autonomous vehicles will arrive in large numbers, Boston Consulting Group states in an April 2015 report entitled Revolution in the Driver’s Seat. Going forward, automakers will add technology that gradually but inexorably takes human drivers out of the equation.
In reality, a working fleet of truly autonomous vehicles demands adoption of a spectacular array of technology. Onboard hardware, software, road infrastructure, communications systems that talk to other vehicles and to infrastructure and more will have to work simultaneously to support a widely deployed AV fleet. The potential payoff, however, is huge. With driver error the cause of 90 percent of all accidents, migrating the locus of control away from error-prone people promises massive commercial and social good.
The fact that both Google and Apple have targeted substantial resources to AV initiatives, along with companies such as Volvo’s partner Ericsson, reflects the rapid datafication of private vehicle transportation and the obvious need for robust data management systems.
Yet the technology, it would seem, is only part of the challenge—perhaps not the most difficult to overcome. As we evolve from increasingly sophisticated and more widely deployed collision avoidance systems toward fully autonomous vehicles, we face a daunting range of administrative and commercial issues.
For instance, where does liability rest when intelligent vehicles are involved in a collision? To what extent do we insure drivers compared to vehicle makers, infrastructure owners, or suppliers who provide systems?
Indeed, it is questionable that vehicle ownership in future remains exactly as it exists today, given the possible opportunities in car sharing that autonomous vehicles engender. Fundamental questions such as who owns vehicles, who pays to insure them, and how is data ownership managed are bound to be in play as insurers adapt to the disruptive impact of intelligent vehicle systems, lower accident incidence, and reduced severity.
Mark Francis, ICBC’s Manager of Driver & Vehicle Licensing and President of the Insurance Institute of BC closely tracks developments in the area of autonomous vehicles.
“We’re looking to the US, and particularly California, to see how they handle a number of issues,” Francis says. “There’s potentially a great deal of complexity around liability, and it will certainly have an impact on insurance for individual drivers,” he says. Insurers will also have to face new pressures because technology and consumer demand for autonomous driving will likely outpace regulations, Francis observes.
At present, regulators in Canada and in BC are looking south over the next year to help them in establishing evaluation protocols and standards for advancing and regulating AV technology. But without even having those standards in place already, regulators will have to move very swiftly over the next five years to provide a legal, commercial and administrative framework for Canada that keeps pace with technology.
It is no longer a question of if autonomous vehicles will hit the road—but rather when. How the industry negotiates Kurzweil’s S Curves will have a critical impact on our expedient arrival in the land of self-driving vehicles.
This article first appeared in BC Broker