Driverless cars will reduce but not eradicate congestion, road deaths, say experts

A self-driving car drives along a road during a presentation in Moscow, Russia, on Aug 16, 2019. Industry observers reckon that driverless cars will be rendered immobile in a mixed vehicle urban environment because they will not move unless it is absolutely safe to do so. PHOTO: REUTERS

SINGAPORE - Autonomous vehicles will not eradicate congestion or road deaths, according to experts speaking at the 26th Intelligent Transport Systems World Congress here.

Dr Mahmood Hikmet, head of research and development at New Zealand driverless shuttle start-up Ohmio, said the communication around autonomous vehicles tends to "over-promise", and one of these promises was that there would be no more congestion and no more fatal accidents once driverless cars become a reality.

He blamed this in part on poor communication between engineers and non-engineers. "The way we communicate technology is not accessible to most people," he said at Tuesday's (Oct 22) session titled "Autonomous vehicles in public transport - separating hype from reality".

He cited the example of trolley problem, referring to how driverless cars are programmed to react to "moral dilemmas" such as who should live or die in an unavoidable accident.

In such a scenario, Dr Hikmet said the vehicles will just stop. "It just brakes," he said. "It's not an engineering problem. It's not rooted in reality, but rooted in what people think reality is."

He said the reason why the trolley problem has been given so much air time is because it makes good headlines.

He added: "The assumption is that something which all the sensors fail to detect appears suddenly in front of the vehicle out of nowhere; and the vehicle has to decide if it's a child or a doll, if it's an elderly person or someone with bad posture or bad skin.

"And if it decides to steer, the vehicle has to gauge the condition of the road, whether it is wet, or there is paint on it.

"It's a hypothetical based on a hypothetical based on a hypothetical. So much so that it's not worth spending engineering time on."

Mr Malcolm Dougherty, national transportation practice lead at American engineering consultancy Michael Baker International, said: "Because of the social challenge of that question, we might have to limit our (autonomous vehicle) speed to 15mph (24kmh). And that won't be very efficient."

Remote video URL

Indeed, industry observers reckon that driverless cars will be rendered immobile in a mixed vehicle urban environment because they will not move unless it is absolutely safe to do so.

Mr Dougherty also raised the possibility of increased congestion if autonomous vehicles are not deployed properly. "We are now concerned about single-occupancy vehicles. The last thing we want is zero-occupancy vehicles running about," he said.

Both Mr Dougherty and Dr Hikmet said the underlying purpose of autonomous vehicles remains the same as that of conventional manned vehicles - to meet increasing demand, enhance service and move people efficiently.

"The goal remains the same: To get to a destination as fast as possible," Dr Hikmet said.

He said autonomous vehicles allow the world to move from a locally optimal state where individual drivers' self interest rules, to a globally optimal state where traffic in the entire road network is optimised.

If this goal is attained, there will be less congestion, less pollution and fewer accidents.

Join ST's Telegram channel and get the latest breaking news delivered to you.