I'd guess that a road like Dalton where weather-related and road condition-related hazards can suddenly appear, would make an exceedingly poor fit for self-driving non-guideway vehicles.
Far better would be vehicles -- self-driving or otherwise -- on some sort of fixed guideway; given that most of the current usage of the road is for traffic solely between the two points-of-interest on either end. Special vehicles could still be used to access the pipeline running alongside at arbitrary points.
A railway, potentially with extra equipment to assist with maintaining traction and negotiating grades, would accomplish this nicely, but is avoided due to its large up-front cost. While self-driving vehicles don't come with such large non-distributed up-front cost, they would be unreliable on such a road without permanent in-vehicle supervision, at which point they're not really making any of the intervening towns that provide driver services superfluous.
Far better would be vehicles -- self-driving or otherwise -- on some sort of fixed guideway; given that most of the current usage of the road is for traffic solely between the two points-of-interest on either end. Special vehicles could still be used to access the pipeline running alongside at arbitrary points.
A railway, potentially with extra equipment to assist with maintaining traction and negotiating grades, would accomplish this nicely, but is avoided due to its large up-front cost. While self-driving vehicles don't come with such large non-distributed up-front cost, they would be unreliable on such a road without permanent in-vehicle supervision, at which point they're not really making any of the intervening towns that provide driver services superfluous.