But as a professional driver for the last few decades, I can assure you that you won't be seeing manless semi's anytime in the near future. The only vehicles they have now cost nearly a million dollars, and you still need a licensed driver in case the unit stops. A computer cannot navigate main roads or side streets. It can't calculate turns or give emergency vehicles right of way. It can't listen to directions by a road crew worker in construction zones. A computer will never be able to back a trailer into a dock. It just isn't feasible.
It takes more than just driving to safely pilot a tractor-trailer. You need instinct and experience. For instance when I see an asshole on the highway weaving in and out of traffic, I have to back down if he gets near me because he's liable to cutoff my safety distance and hit the brakes. A computer could never calculate assholes. If I hear something fall over in the trailer, I have the ability to stop and see what happened and correct the problem. A manless truck would keep on going which would be dangerous because at times, we haul carts that are on wheels. A cart that broke free of load locks or straps could easily bust through the back doors of the trailer. And even if there was some miraculous way to address those concerns, there is an insurance issue to consider. Insurance premiums are huge for trucks unlike cars. Insurance would be unaffordable for a manless truck. But I digress:
I would disagree completely. Not only will driverless vehicles take over transportation in general over the next decade or two IMO but they will do a FAR better job. A driverless vehicle makes up a TON of productivity because it does not sleep and I guarantee that it very well can take that 'asshole' into account and even more - it can do so faster while considering millions of different possibilities in the time that it takes a person to blink. Backing up will be a simple task - I do not see why you think that a computer could not take basic physics into account to back a trailer up. That is an extremely simple task for a machine and many machines do far more precise movements with much more complex maneuvers. There are many other tasks that are far more complected than that.
They certainly do not cost a million each and, more importantly, THEY ARE ALREADY HERE in a commercial capacity:
Google's Waymo is launching competely driverless taxis this year
Every major automaker is unveiling a concept car this year as well that literally has no controls. No steering wheel or pedals to speak of.
The reason that drivers are required so far is not because of the technological barriers - google has already proven the concept by driving millions of miles with a single fender bender - but because regulation is always slow to respond to quickly moving industries. Now that the law is getting on board, you will see things move faster and the tech will become cheaper as it grows.
But it's not going to happen in our lifetime, trust me.
Driverless vehicles have been on the table for over two decades, yet when put to the real test, they fail all the time.
But even if they ever could perfect the driverless cars, it would easily take at least another 20 years for trucks to follow.
Uber suspends self-driving car program after Arizona crash
Here's what happens when a self-driving Uber fails
It reminds me when I got my first computer and I bought a dummy's book. The author advised never buy new software for your computer. Don't buy version 4.0, because 4.0 doesn't have the bugs worked out yet. Wait a while for version 4.2 or 4.4 to come out, because those are the versions with the necessary adjustments and repairs.
Software works great when you pass it around to a few thousand people. When you pass it out to millions, that's when the problems arise.
So even when cars (or trucks) come out, we won't know the results until millions of people have them. In the meantime, trucks can cause serious damage and multiple deaths in a severe accident. The insurance to cover those vehicles will probably cost more than laying off ten drivers.
I wouldn't get too complacent. Self driving vehicles will be twice as good in a year, four times as good in two years, eight times as good in three years...
If that were true, we would have had manless vehicles fifteen years ago.
I think that if they could possibly perfect the driverless car, it would save lives. In spite of our great success in the reduction of impaired drivers on the road, there are still millions more out there. In the future, if a person wants to go to a party or a bar, they could drive their own vehicle their destination, punch in the coordinates, and simply get in their car at the end of the night and let the vehicle take them home.
As for trucks, it's a different ball game. They have been installing more and more technology in trucks the last ten years or so, mostly due to pollution crap. Now these things break down all the time and it gets costly because it's all electronic stuff. I have the newest tractor in the fleet. It has 12,000 miles on it. It's already been in the repair shop twice, and once somebody had to come out to get it started. 12,000 miles for a truck is like 1,000 miles for a car. Imagine how pissed off you would be if you had to take your brand new car in twice in under a thousand miles.
The experimental vehicles they are screwing around with now can only do one thing, and that is go straight on a highway. Technology? Hell, we can't even find a GPS that can lead you to your stop using truck routes only. Often they F-up and take you down a truck restricted road, or lead you down a road that has a low bridge that you can't pass. I don't even use the damn things. Too untrustworthy.
So we are a lifetime away from total manless trucks; not that I'm worried, I'll be retired in five years from now. But I just don't see it with the experience that I have. There are just too many human calculations that have to be made on a daily basis to pilot a tractor-trailer that computers could never do. And remember, automation is only good if it's cheaper and just as dependable as human labor.