Connected and autonomous vehicles

I’m becoming exponentially more skeptical of the value of connected/autonomous vehicles. I’ve known L5 automation is still a long way off, but some of the reports suggests many companies are still working on L3 automation. And there is zero evidence that the developers are working on anything system able to perceive pedestrians, implying that the cars will never be able to operate in a pedestrian rich center city environment.

Operating in ideal conditions (peak hour traffic, well-striped lanes, controlled access, no pedestrians/cyclists) could happen tomorrow. But that exposes CAV to their own ‘last mile problem’, on both ends of an auto trip, where such conditions are absent.

Starting with limited access highways, with ‘connected’ pylons to serve as a High Accuracy Repeater Network for GPS navigation for expressways and other divided highways is easy, and probably useful, as they represent the majority of VMT. But they are an extreme minority of lane miles, and extending the network beyond that would be problematic: road quality falls off quickly—striping is incomplete/damaged/incorrect, curb/gutter are sometimes absent, curb-cuts proliferate, and pavement quality declines.