How long until truck drivers disappear?
We have already gotten rid of the factory worker and that has led to withering away of Middle America. Removing the truck driver would be a huge nail in the coffin of Middle America. (If lab grown meat + vertical farming takes off, we can potentially drastically reduce the number of farmers in this country.)
I am all for these technical innovations, but we must be aware that this can dramatically change the economic/political/social landscape. Radical regressive ideologies spring out of such environments.
I think dock workers used to unload ships and my understanding is, that while not as extreme, they have done something similar. (Not that I see this as the best approach....).
A majority of farmland and ranches are already large corporations, so I doubt lab grown meat has that significant of an effect.
Short-haul, maybe never.
Energy use from Human QoL stuff would be very very low in a semi compare to the energy use for pulling.
For instance, the tractor and trailer, unladen weighs 25,000 to 40,000 lbs, with a total road weight of 50-100k lbs. [1] A whole passenger car with amenities can weigh less than 3000 pounds. Call 1/5th of that QoL stuff, for 600lbs. Changes in automobile weight matter much more because automobiles stop and go much more often.
So the human QoL stuff probably doesn't weigh more 1% of the road weight. But the kicker is, it doesn't even matter too much if the QoL stuff weighs 10%, because energy use scales mostly by air resistance, not weight. [2]
Most of the QoL energy waste is found in idling semis at truck stops or in traffic, so this could definitely be reduced by Auto-Semis [3]
The eventual savings from Auto-Semis will be in reduced wages, and possibly in reduced fuel usage - but the reduced fuel consumption would come from driving slower, as you don't have to pay a driver for their time, and an Auto Semi can drive all night.
I'm sure logistic companies would love to save this much money, but the capital expenditure required will be huge. Some research from passenger self driving cars would be pertinent, and probably most sensor devices, but they would be put to slightly different uses - you want a longer range view in a truck for braking distance and crash avoidance.
The reason I put the estimate so far out is liability and regulation and the cap ex mentioned above. Some classes of load will probably never be pulled by auto-semi: dangerous, time Sensitive, some luxury goods.
Auto Semis could also do some interesting things like draft (tailgate) safely and form a "train" of trucks. You could have drone drivers set up who monitor 3 or 4 trucks and take over if needed.
1. http://www.thetruckersreport.com/truckingindustryforum/threa...
2. http://www.theicct.org/sites/default/files/publications/AERO... - Page 5, PDF
As Ucho says briefly, and I say below, the biggest economic difference between human operated and self driving semis are:
* Tonnage / Time - the Auto Semi can drive all the time, no rest breaks.
* Fuel usage - Auto Semi can drive to an algorithm, not just what "feels best", and the Auto Semi can drive slower, because you aren't paying a driver for their time. Auto Semis can "platoon" for fuel savings from drafting, among other ways to achieve savings.
* Auto Semis will probably always have a human minder, at first in the cab, for experimental trucks, and later in a remote operator center (think drones)
* Insurance - at some point in the future insurance for Auto Semis will be much less human driven trucks.
* As soon as the break even point for the investment of sensors is about 1 year (Earnings from hauling, Fuel costs, and wages), the big logistics companies will BEGIN to replace their fleets. I expect this in 5-7 years, based on no hard facts what-so-ever.
1. http://money.cnn.com/2015/05/06/autos/self-driving-truck/