Ford plans to be making self-driving cars for the mass market in five years’ time. Having been caught behind most other major car manufacturers in the development of the electric car — a viable prospect — Ford are now trying to catch up with driverless cars.
Whereas electric cars will require all sorts of other (attainable) research and development — laser radar sensors and long-life batteries, for example — self-driving cars depend essentially on a one-stop strategy — the successful development of error-free software, General Artificial Intelligence (GAI), that’s as versatile as the human sort. So far, this has been impossible.
It is the sudden flash of unexpected data which causes driverless cars to fail. On a motorway or a largely featureless countryside road with little data about them in the car’s computer memory such a data input can cause the car to make an inappropriate — indeed dangerous — decision sometimes.
In a city street with multiple visual points of reference already stored in the car’s computer memory, any unusual flash of new data can be immediately compared with a preponderance of pre-existing data. Thus a more ‘weighted’ relevant decision can be arrived at — rather as a committee usually takes wiser decisions than individuals.
Therefore, driverless cars could well be suitable within cities — driverless taxis or uber-cars, for example — but I can’t see them ever being popular for the daily commute from the suburbs or for freight traffic on a busy mororway.