"Autonomous cars won't prevent 90% of accidents"

google autonomous car
google autonomous car

Road safety expert Dr. Ian Noy sees the hype and expectations about driverless cars as greatly exaggerated.

Theres no factual basis to the claim that autonomous cars will prevent 90% of traffic accidents, said Israeli-born Dr. Ian Noy, who has filled a number of senior roles related to road safety in Canadian governments. He recently visited Israel as a guest of the annual Or Yarok road safety convention.

The sterling reputation of autonomous vehicles, says Noy, is based on global statistics which claims human error as the cause of 90% of accidents. Any interested party and there are many who believe it will promote their businesses can rush to the conclusion that the autonomous vehicle, which takes out the human element, will avoid all these incidents.

If only that were true, says Noy with a sigh, But the truth is human error is the default explanation for every accident. Its what police officers and accident investigators write on their reports because they cannot prove the cause, not because facts show a misjudgment by the driver.

Dr. Noy received his professional accreditation as a civil engineer at the end of the 1960s in Toronto, Canada. After a successful career focusing on roles in the intersection between safety, transportation, and civil engineering, he relocated a decade ago to Massachusetts.

There he was offered his current position as VP of a Boston research center Liberty Mutual Research Institute for Safety which is funded by the big insurance companies as their contribution to social advance. The center funds research in transportation and road safety, which are freely distributed to the public.

The evolution of road safety began more than 60 years ago. It started with research and development of protective auto components such as seat belts and airbags. The second phase emerged with the development of accessories and secondary systems built to actively avoid accidents, such as the new early collision warning systems and speed alerts.

The current stage in the fields development is focused on the drivers, their reactions, and their conduct.

One way or another, says Noy, Accidents, which have terrible consequences, are extreme cases. Autonomous vehicles have certain advantages over drivers because they do not tire, or become inebriated and they are not distracted by problems with a partner or a supervisor, but they face many challenges even before we touch on the safety issue, challenges which relate to their integration into the flow of traffic.

Can you give an example?

The technology which controls the actual driving has not yet matured. Their test drives are conducted under ideal conditions on straight-aways, on highways, when only other cars are on the road, when there are no pedestrians or bicycles. Even in a utopian scenario, in which all the autonomous cars are directed by a central, coordinating system, it will not be able to control the movements of pedestrians and light transportation vehicles which are being increasingly used making autonomous driving in urban areas far more complex.

What options has the industry developed for urban environments?

The autonomous vehicles we are discussing are not always autonomous. They ask the passenger to take over when the systems are not confident in the safety of the automated driving. Essentially, these are hybrid autonomous vehicles, similar to airplanes, which for the most part fly using their instruments. But no one is suggesting we let airplanes take off without a pilot, right?

Then what do we gain?

That is the essential irony of innovation, that we expect it will improve current conditions but every once in a while the innovation gives rise to new problems. The autonomous car will assist with the easy part of the drive along the highway and when you enter the city, it will ask the passenger to revert to being a driver.

But that is its own problem: the driver now has to take control of a vehicle with complicated technologies and be adept in using systems and components that do not exist in traditional cars.

Another issue is the integration of the autonomous car with a large transportation system which controls a fleet of vehicles and interfaces with other systems including stop lights and smart parking structures. The driver will need to be aware and take into consideration a host of other inputs, not only what they can see.

Instead of simplifying, we may end up complicating both the act of driving and our way of life. After all, the best autonomous vehicle already exists the taxi cab.

Like credit cards

The entry of tech giants Apple and Google reinvigorated the traditional players in the car industry as they suddenly identified a new threat. The problem lies in the closed garden approach as each company is developing its own autonomous car and its research is not shared with any other firms.

The behavior of an autonomous vehicle whether manufactured by Apple, Google, Mercedes, or Volvo is not shared with autonomous vehicles from other makes. The lack of coordination can hurt the flow of traffic and the overall safety.

For example, an autonomous vehicle may be programmed to wait at an intersection until all other vehicles come to a full stop. In many cases, cars do not come to a complete standstill for long enough to be registered by the system, and the intersection might become a bottleneck. The autonomous cars may wait there forever, he says, And thats an example of a minor problem.

Another troubling scenario involves the decision making of an autonomous vehicle which identifies a road hazard dangerous to its passengers but also recognizes that swerving would lead to hitting an adjacent vehicle. What will the computer do? Will it protect its passengers? Would it risk harming the passengers in the next lane?

To solve these and other issues, Noy suggests learning from other industries in which the major players are in competition but also cooperate in a manner that prevents industrial espionage while allowing the actors to profit. He cites air travel and banking as two examples. Because the banks coordinate, he says, you can use a small plastic card to buy almost anything anywhere on the planet. Businesses worldwide, credit card companies, and online retails all comply with the guidelines set by Visa.

In that industry it was easy, because everyone wants to make a sale, he explains. And all airplanes are free to take off and land from any airport, because they follow a shared lingo and industry standards. That needs to happen in the autonomous vehicle world. The lack of coordination can lead to immense hitches on the roads and place passengers in danger. For autonomous vehicles to start operating routinely we must set identical rules for the entire industry. Otherwise, it wont happen.

When will we actually see an autonomous fleet on the streets?

I do not know when, but I am certain it will enter the market significantly after Europe and the Far East. The reason comes down to liability. When an autonomous vehicle is involved in an accident, its difficult to pin down blame and determine who must pay: the driver, the manufacturer, or the operator. The well-oiled litigation industry in the US will delay their introduction, concludes Noy.

And as a senior official in a non-profit organization funded by the largest insurers in the world, his outlook is invaluable in this debate.

Published by Globes [online], Israel business news - www.globes-online.com - on April 8, 2016

Copyright of Globes Publisher Itonut (1983) Ltd. 2016

google autonomous car
google autonomous car
Twitter Facebook Linkedin RSS Newsletters Israel Business Conference 2018