Robot Cars are Coming!

The biggest obstacle I can see to robot cars is being able to share the roadways with non-robot vehicles. Failure of computer systems will always be a factor to consider, since nothing is 100% perfect. Then there is the human factor. Can a human ever feel secure with a non-human at the wheel? Is he able to put his confidence 100% behind a robot? I think not.

The time is coming when robot computer controlled cars will be more reliable than the human controlled ones. There may even be a movement to remove the human controlled ones from the roads.

How long before some manager "saves money" by re-using computer code intended for a "Smart"-sized car to control, say, an 18-wheel fuel tank truck? See http://en.wikipedia.org/wiki/Ariane_5_Flight_501 for some idea of what might happen. I don't doubt that, given adequate specialized infrastructure, robot cars will become more reliable than humans (though only while operating within that special controlled territory) - but how could we bridge the two or three - at least - decades during which both human- and computer-controlled vehicles will have to share the roads? The alternative to providing infrastructure - possibly as simple as painted lines, low-power signal cables, etc. - involves the requirement that far more sensors, more computing power, and more [inherently unreliable] software be installed into EVERY vehicle. And given the need for such infrastructure, how do we deal with the "last mile" issue (analogous to the cable service version: http://en.wikipedia.org/wiki/Last_mile) between the major highway and the homeowner's garage? Remember too that the dependence upon a computer to "drive" the car will mean that the owner/passenger will have far less practice driving than is today's norm, so the problem will get worse for a long time after such vehicles become available.

Even if we could design, test, debug, and install special guidance provisions on every major highway, and on every city street, do it quickly and at low cost, we STILL need quite a bit of electronic equipment on every "robot" vehicle, in order to deal with pedestrians, bicyclists, fallen trees or rocks, debris, large animals, failed roadway sections, and those pesky human-controlled vehicles . . . and we still need the human-interface operating controls to handle rural roads, farm lanes, alleys, driveways, extreme weather conditions, and so on, so there is no cost-saving potential in that area. The need for both systems and the overrides and protocols will make this portion MORE costly, in fact. Cost WILL be a limiting factor to introducing such cars/trucks/busses/motorcyles. A politician or bureaucrat attempting to initiate the changeover by law would be out of a job rather quickly, long before it could take effect.