This is exactly the type of article I like to see on HN. It presents a simple solution to an obvious problem that I hadn't even really recognized, and it reminds us to keep looking at things in different ways.
One of the most vexing aspects of traveling between mainland China and Hong Kong is the car travel: People in the former drive on the right side of the road; people in the latter drive on the left (a vestige of the British empire).
So to quell confusion at the border and, more importantly, to keep cars from smashing into each other, the Dutch firm NL Architects proposed a brilliant, simple solution, the Flipper bridge.
The most standard design (called A4 in the wikipedia article) has the same amount of traffic lights as this design. There is no driving on the left, no weaving, and there are no left turns with oncoming traffic.
Access onto the highway is a right turn for all directions. Access from the highway is light-controlled right or left turn. The highway has more capacity than the intersecting road, so while dumping traffic from the road onto the highway at a maximum rate using the easy right turns is fine, the light at the exit off the highway serves as rate limiter if necessary so that the road is not overwhelmed with traffic.
To be sure, those interchanges are the most pleasant to drive through, but at a cost of land. The interchange design presented in the NPR article is much more feasible in urban centers where you might be constrained by existing structures. The accel/decel ramps can run parallel to the freeway, leading to nearly-right angle turns at street level. (It's not drawn this way in the article, but it doesn't take much imagination to picture it.)
That's definitely a good point, but in this particular case the picture posted elsewhere in the comments by raquo [1] looks like it could fit or almost-fit a parclo (possibly in a slightly elongated/compressed arrangement). The more hippie side of me is tempted to suggest that if you can't afford the land, you shouldn't be building an expressway there :)
Ditto on space constraints. Note also that in urban centers existing structures also include underground electricity, water supply, gas, phone lines, etc. which are usually very costly to relocate.
It seems that this will still suffer the weaving problem that plagues cloverleaf intersections - note how on the bridge cars first enter the road (from the freeway) and then shortly afterward, leave.
Perhaps this could be mitigated by placing stoplights on the offramps (which can be conveniently anti-synchronized with the lights controlling the bridge entrance), but I imagine space on the offramp could be a huge issue.
The main idea is that it's a much cheaper upgrade to the diverging diamond from a standard diamond interchange with left hand turn lights, but it's still a fairly substantial upgrade for traffic flow. The cloverleaf eliminates all the lights, but requires a lot more construction and takes up way more land.
Seems like a lot of lights, compared to a roundabout. I don't get why people don't like roundabouts, they seem to work quite well, at least until traffic backs up (at which point nothing really works). They also avoid the left (or right, where they are most common), but typically with fewer lights. The second light in this system seems like it will really slow things down, but maybe it works. I'd love to see video footage of it in action during a busy period.
I have a sort of visceral negative reaction to roundabouts, but I think because most of the ones I've seen in Europe, when the road is more than one lane wide in each direction, seem to end up with a crazy ad-hoc-lanes thing going on in them, where it's as if you drove onto a huge plaza with no lane markings, where everyone is driving in all sorts of criss-crossing diagonal directions trying to get from their entrance road to their exit road across multiple lanes of traffic. I imagine some countries have more orderly versions, but that's how the Paris ones all felt to me.
I approve of one-lane roundabouts and wish America would replace all its four-way stop signs with them. Four-way stop signs make no sense and just make everything painful, especially for cyclists.
Two-lane roundabouts are pretty dangerous, though -- while I've never had an accident in one I've always suspected that I might. And anything more than two just seems to be crazy.
Well, the Arc de Triomphe is not exactly a /good/ example.
Your point is valid, but I (and others more qualified) argued this ends up being safer. People's mental model changes when they get to a roundabout, they sort of switch on their brain a bit more. I have in my head, but can't find a reference, that roundabouts are actually surprisingly safe.
I think the biggest problem with such a design is simply people not being used to it. Once you're used to occasionally being on the left side of the road for thirty seconds twice a day, there's no downside. The first took a long time, the next few will come slowly, and then they'll become just part of the terrain like cloverleafs (bleh). I also expect their acceptance will vary geographically, like roundabouts (still bleh for me, YMMV).
I live in Springfield, and so far it actually seems like the most dangerous parts are the right hand turns from MO-13 on to I-44 (step 6 on the interactive diagram). A lot of people forget that although they have bypassed the traffic signal they still have to yield to the oncoming traffic from the opposite side of MO-13 that they would typically be looking at to check right of way.
I think the "newness" of it all could be alleviated simply by posting traffic officers at both intersections for the first month or so of usage. People are always more alert when they see a cop a round and they will be there on the spot if any issues arise.
I found a video that helped me to visualize this design better than I could from the article. It is a 4 minute long animation of the interchange in action from the Missouri Department of Transportation found on this page: http://www.modot.mo.gov/springfield/major_projects/Greene/I-....
Am I the only one that's bothered by the left-lane merges? That's really difficult to implement, esp. in the way it's drawn in the diagram in TFA (no lengthy merger lane).
In the left lane, people go fast. It's dangerous and difficult to merge from an exit ramp onto the left lane.
The roads that merge both have only one lane. The one that comes from the right has a yield sign. It also has a rather sharp bend just before the merge so 1) the cars are unlikely to go at high speeds, and 2) the drivers can easily see each other before the merge, so it's actually good design.
I believe ComputerGuru was talking about the merge into the left lane of the intersecting road while exiting the highway and turning left. It's not the most intuitive of arrangements, especially when combined with the left-hand exit following shortly. Hopefully the drivers on the intersecting road will be sufficiently surprised by the arrangement to break the usual expectation of the left lane being a through lane with higher speed.
This looks like a good idea. Picture your typical urban freeway interchange with left turn lanes to get onto the freeway. This design eliminates the left turns by moving each direction to the left on the bridge: English on the bridge and American on the approaches. There is a diagram that makes it all clear.
I agree, looks like a pretty good idea. Any type of change like this always gets negative opinions, same thing happened when they put roundabouts in and now everyone loves them and want more of them.
Love this kind of stuff. It can be trippy driving through these things though.
How long before traffic systems start broadcasting the schedule for the next minute or two so cars can plan ahead and always get there when the lights green?
The article ran in November of last year, so the interchange has (presumably) been used for half a year now. Surely Missouri's been collecting data about how the new system's been doing - I'd love to see it.
I'm sure I understand why the stop light/intersections need to be there at all. Why can't the orange road pass over the red road on one side and then pass under on the other side of the underpass?
Building overpasses isn't exactly cheap and the article said no bridge construction was necessary when putting this layout up. It can definitely be done if the traffic level warrants shelling out extra for not having two traffic lights, but once you're going for a three-level interchange there are a lot of existing designs out there already (and some don't have the entrance-exit weaving pattern).
Additional bridges not only increase the cost of the interchange, but also its size. To allow for a 5 m vertical clearance for the bridge, you will also need approach and exit ramps. If your maximum allowed design slope is 5%, you will need 100 m ramps at each side of the bridge, increasing the size of the interchange by 200 m.
