Tech torque: Self-driving cars

Back in the mid-1980s, so-called ‘flying cars’ were being talked up as the way of the future.

They were going to make roads obsolete and traffic jams a thing of the past. 30 years later we are still waiting for them to take off.

Now driverless, or ‘autonomous’, cars are making all the headlines and, if you believe the hype, are about to become commonplace on our roads. A recently released research paper by commercial UK ‘think tank’ Juniper Research predicts autonomous cars will be widely adopted by consumers in five years, and in ten years there will be 20 million self-driving cars on the world’s roads.

I guess there’s nothing like positive thinking, but it seems this is very much a pipe dream. Lots of things will stand in the way of the adoption of autonomous cars. Some of these are obvious:

• For people who don’t wish to drive, there are alternative forms of transport including taxis, buses, trams and trains. Perhaps more significant will be the rise of ride-sharing, phone-app-driven services like Uber, as they provide an easy and convenient way to get about without having to drive yourself or even own a car.
• Autonomous cars won’t be cheap. They will need a whole raft of high-end technology such as radar, cameras and proximity sensors, not to mention sophisticated control software as well as robotics to control the steering, brakes and throttle.
• Conventional vehicles come in myriad types such as sports cars, hatchbacks, sedans, wagons, people movers, vans, utes and four-wheel drives. There is no way that autonomous cars will be offered in such a wide range. They will probably be restricted to small, city runabouts that will potentially satisfy just one segment of the new-car market. That alone will limit their market penetration, at least in the short and middle term.
• Autonomous cars will present a difficult challenge for lawmakers and vehicle insurers. Who is to blame if an autonomous car runs over a pedestrian? Will it fall back on the carmaker, the car owner or the software developer? If an autonomous car is faced with the dilemma of hitting a pedestrian or veering off to the side and into another car, what decision will it make and who is to blame if it makes the ‘wrong’ decision?
• Driving a car in heavy traffic on a multi-lane expressway, or in wet or slippery conditions, or down a winding and unsealed country road with varying levels of grip, or in any other demanding driving environment requires a driver to make complex and rapid decisions based on experience. This will be difficult to replicate in the central control unit, or ‘brain’, of an autonomous car.
  Even something as simple as a plastic bag blowing in front of a car, which a human driver will quickly perceive as a non-threat, may trigger hard braking from an autonomous car, potentially causing a rear-end accident.
  Likewise, when a pedestrian is waiting to cross a road, or is already crossing a road, a human can quickly access whether the pedestrian has seen the vehicle. If a pedestrian is looking the other way, a human will be alert to a potential problem. It’s hard to see how an autonomous car will be able to achieve the same level of sophisticated decision making.
• Autonomous cars will need their central control unit, or ‘brain’, to be 100 per cent reliable and ‘freeze-free’, something that has not been achieved with PCs, laptops, tablets, or even smartphones to date.
• Advocates of autonomous cars claim that they will be able to ‘talk’ to each other to help traffic flow. Nice idea, but only if they all have the same operating system with current updates. And, either way, autonomous cars will always have to share the road with conventional vehicles and put up with the foibles and often inconsistent driving of other road users. Experienced drivers can readily spot an ‘idiot’ driver and give them a wide berth, but will autonomous cars be that savvy?
• Google is currently leading the way in autonomous car development, but it has zilch experience as a carmaker. Moreover, Google cars will rely on Google Maps for navigation, a critical technology for any autonomous car. Google Maps and its associated sat-nav is poor at the best of times. This doesn’t bode well for a Google car being able to find its way, especially away from major population centres.

All of this doesn’t mean that autonomous cars or other autonomous vehicles won’t feature on our roads sometime in the future. The technology will certainly be useful on something like a bus that follows a fixed route, or in other simple applications where the variables are limited. As for it becoming widespread across all vehicles types, that’s something that is, at best, decades away.

GOOD
In a utopian society, the self-driving car would rid the world of traffic congestion. Who actually believes that’ll happen, though?

BAD
RELIABILITY. No matter how good your PC or phone is, sometimes it’s gonna freeze. Same goes for the self-driving car.

Follow The Leader

Many of the technologies that will feature in autonomous cars are already in use and are becoming increasingly common. This includes adaptive cruise control, which automatically limits your vehicle’s speed to the speed of the vehicle in front, and will only allow your chosen ‘cruise’ speed if the lane ahead is clear.

Adaptive cruise control relies on forward-facing radar, also central to the automatic braking systems on many of today’s newer cars. With automatic braking, the car will apply the brakes if it perceives a threat that the driver hasn’t reacted to.
Other current technologies that will be used on autonomous cars (but in a different role, given there won’t be a driver to alert) include blind-spot monitoring and lane-departure monitoring.

On current vehicles all of these technologies have some use, but none are perfect and some are even flawed, so there’s some work to do to perfect these basic autonomous vehicle features.