Well, for one, the idea that the leftmost lane is reserved for speciality traffic is something that you guys have in the US but that hasn't permeated to a lot of other countries. Highway bus lanes are generally on the right over here. In almost every country the left lane is a passing lane, so it doesn't hamper your idea, just something to note.

What people do often forget is that AVs are heavily sight-based, and will not risk something they can't see. So while I can see the first part of the ramp being shorter, any yaw or negative pitch change will have to be taken significantly slower than we would take them. The benefit of not having to cross all lanes is obviously a great one, but the fact that many urban highways don't have a significant median might make it really difficult to construct.

My idea was not at all focused on highways, really. Here's my elevator pitch now - I invite anyone to gimme some feedback because it's a relatively early idea. (kleinbl00?) The context is European city centers.

Autonomous vehicles are coming, but their biggest challenge will be to conquer the densest urban areas. Precisely the places that most people will want to go, are also the most difficult to traverse on public roads. Some estimations put dense urban autonomous vehicles at at least 40, 50 years into the future. Infrastructure changes, most notably dedicated AV lanes, will likely be a necessary step in the transition to autonomous vehicles.

Autonomous vehicles present a threat to public transport, so it is wise to design autonomous infrastructure in such a way that it also strengthens the public transport network instead of only competing with it. At the same time, due to the high uncertainty of the precise development path of AVs, it is important that any infrastructural investments are already useful now, instead of waiting for the revolution to happen.

Thus: the idea to build bus lanes or expand the bus lane network in cities so that they connect the most important locations with highways. In the current PT paradigm, this would allow regional buses to easily drive on into cities and compete with cars more easily. In that sense, it's a 'no regret' investment. In a highway-AV-only scenario, the highway location could be an important transfer hub for the last miles into the city. In an AV-on-separated-lanes scenario, the bus lanes provide the AVs with the separated environment they need to maintain high speeds in dense urban areas. And because they enhance already existant high-quality PT with intermodal transfer hubs, they can significantly strengthen the PT network.

You know why I have an iPad? This is why: