18 July 2016

Autonomous Cars : Part 1

Is this sort of image creating an unrealistic expectation?

There is much hype around cars that can drive themselves. Those who are funding the development of them say they will create a safer environment on the road. One publication was so upset that a car on autopilot recently led to the death of the driver, it lashed out at "the extreme fallibility of human drivers" as if to minimise the artificial intelligence's failure. So what can we make of it all?

1) Does human driving reflect extreme fallibility? The fact that humans of all sorts take to the roads, then fatalities per km are remarkably low. If it is such a concern, drivers should be retested at sometime in their lives beyond the first getting of a licence. I have only had one driving test all my life.

2) Are autonomous cars going to reduce fatalities? By taking the driving off the driver while expecting him/her to keep as alert as a driving human is unreasonable. Added to that a person who spends little time driving will lose skill and sharpness. In addition, can artificial intelligence ever have the ability to read and anticipate every scenario?

3) So what is autonomous driving's raison d'etre? I see it best in motorway situations, taking over in heavy traffic and relieving stress. It would also reduce breaking ripples. There is not that much for the computer to do and the driver will be kept fresher.

In summary, I wonder if some are overstating how far technology has gone and can go in the foreseeable future. Funding may be reduced if the progress per dollar spent was more realistically costed. The recent death mentioned at the outset could be because a human was relying too much on the auto pilot. Perhaps the rhetoric coming out of the media about autonomous control of cars has mislead people to expect too much.

The picture accompanying this article shows how far some may think autonomous driving has gone or can go. In reality, the scenario pictured could take many years to happen or may never be achieved. Trying to get technology (created by humans anyway) to mimic human abilities is impossible. Computers are good at number crunching and taking on tedious jobs but a high level of anticipation, understanding and reaction is best left to the human driver.

Part 2 can be found by clicking here.