At was speed does the automated system pay for itself if speed vs human is a reasonable trade off to consider?
* Not considering the cost of doing the research and the cost of the self-driving truck.
For the use case mentioned in the article the human driven option is going to be cheaper because the human driven truck goes faster. Going 10km/h results in basically the same throughput as having two trucks (yes I know there's other "per trip" overhead, that's not the point).
There are some entertaining (if biased) post-mortem articles on that startup: https://www.fastcompany.com/3028159/a-broken-place-better-pl...
My take is that at 30km/h they would make a very compelling case - since that's the minimum highway speed in some countries.
Add a "package truck" wagon to the back of your train, the truck can book a place, RORO. Then continue on local roads at other end.
In NL it is 70 kph, and if you go that fast you are a danger to other drivers. What country has 30 km/h as the minimum highway speed? And what do the highways there look like? Do people actually go that fast on the highways or is that the speed that won't get you ticketed if you drive on the shoulder with an SMV?
The route is fairly short so speed is probably not an issue: https://goo.gl/maps/9yRtQV88NEzDEcePA
The route is described a bit better in a Swedish article I found: https://teknikensvarld.se/forarlos-lastbil-far-tillstand-att...
A giant leap for regulators, perhaps, since it's technically a public road, but looks like the standard move around company facility use case proposed for autonomous vehicles rather than anything comparable with transit haul or a taxi service.