A more efficient SPUI?

[tradephoric]

Your standard SPUI has 3-signal phases (cross-street left turns, cross-street through, off-ramp left turns).  Here's a SPUI that has the cross-street left-turns set back (it's technically not a SPUI, but I don't know what you would call it).  This allows the off-ramp left turns and the on-ramp left turns to cycle together, essentially eliminating one of the signal phases.

https://www.google.com/maps/@38.4415584,-90.3799409,485m/data=!3m1!1e3

[cl94]

Your standard SPUI has 3-signal phases (cross-street left turns, cross-street through, off-ramp left turns).  Here's a SPUI that has the cross-street left-turns set back (it's technically not a SPUI, but I don't know what you would call it).  This allows the off-ramp left turns and the on-ramp left turns to cycle together, essentially eliminating one of the signal phases.

https://www.google.com/maps/@38.4415584,-90.3799409,485m/data=!3m1!1e3

While they could in theory operate independently, the interchange would fail if left turn volumes are high, as excessive queues would block the exit ramps.

[johndoe]

I've seen this called a "three-point diamond":
http://waset.org/publications/9999613/an-analysis-of-new-service-interchange-designs

I know only of that one in real life.

[tradephoric]

Nice link johndoe.  The author came to some interesting conclusions:

VI.CONCLUSIONS
The results show that there is no superior design in all measures in the interchange category investigated. The standard diamond, the most common service interchange design at this time, has some good features like excellent service for crossing pedestrians, but has generally poor capacity and signal progression. The DCD, currently the popular choice as a diamond replacement, is a fine design, and is as good as or better than most other designs on most measures examined. However, bridge width was the only measure for which the DCD was not beaten by some other design. On the other hand, the DCD performed relatively poorly for capacity and for serving pedestrians. Each of the other designs had their good features and poor features.  Generally, the MUT with slip ramps was best for capacity and conflict points, the tight diamond was best for ROW, the spread diamond was best for unusual maneuvers, and the MUT was best for pedestrians. The single-point, DLT, contraflow left and three-point all had some positive attributes as well. The superstreet design had the best signal progression, was generally good for capacity, and had the fewest negative attributes of all interchanges tested.

[peterj920]

https://www.google.com/maps/place/Rice+St,+St+Paul,+MN/@45.0063461,-93.1058046,18z/data=!4m2!3m1!1s0x52b32a0ae698462d:0xa7cfaf5bebcba1b0

Here's an interesting one at MN 36 and Rice St in North St. Paul.  The eastbound exit and entrance ramps flyover MN 36 to form a standard signalized intersection at the interchange.

[cl94]

https://www.google.com/maps/place/Rice+St,+St+Paul,+MN/@45.0063461,-93.1058046,18z/data=!4m2!3m1!1s0x52b32a0ae698462d:0xa7cfaf5bebcba1b0

Here's an interesting one at MN 36 and Rice St in North St. Paul.  The eastbound exit and entrance ramps flyover MN 36 to form a standard signalized intersection at the interchange.

This is probably the better one. Depending on terrain/conditions, it may or may not be more efficient. Still only one signal.

[johndoe]

People refer to that as an "offset single point".  (Some say spdi instead of spui since they're not always "u" (urban) , but are always "d" (diamond)). It seems more OSPDI (compared to 3PD) can be found built/studied online:
http://www.fhueng.com/services/civil-engineering-design/interchange-design/i-225-alameda-avenue/
https://gettingaroundmpls.wordpress.com/2011/03/10/extra-extra-money-wasted-on-extravagant-highway-project/

[jeffandnicole]

Nice link johndoe.  The author came to some interesting conclusions:

VI.CONCLUSIONS
The results show that there is no superior design in all measures in the interchange category investigated. The standard diamond, the most common service interchange design at this time, has some good features like excellent service for crossing pedestrians, but has generally poor capacity and signal progression. The DCD, currently the popular choice as a diamond replacement, is a fine design, and is as good as or better than most other designs on most measures examined. However, bridge width was the only measure for which the DCD was not beaten by some other design. On the other hand, the DCD performed relatively poorly for capacity and for serving pedestrians. Each of the other designs had their good features and poor features.  Generally, the MUT with slip ramps was best for capacity and conflict points, the tight diamond was best for ROW, the spread diamond was best for unusual maneuvers, and the MUT was best for pedestrians. The single-point, DLT, contraflow left and three-point all had some positive attributes as well. The superstreet design had the best signal progression, was generally good for capacity, and had the fewest negative attributes of all interchanges tested.

Except earlier in the article, the author appears to admit this design is one he created and isn't actually in existence, so there's no real world application and verification it's actually a superior design. Furthermore, note the use of *but it should be efficient*, which means the author has no clue if it would work:

If the interchange site in question has room along the surface street for u-turn crossovers, a superstreet interchange as created by the author and as shown in Fig. 4 (c) may be an efficient design. A left turn from the freeway to the surface street is made by first turning right and then making a u-turn at a median crossover. A superstreet interchange has six two-phase signals–one at each median crossover and one at each off-ramp terminal--but they should be efficient because they are independent by direction and therefore allow perfect progression in both directions at any speed and any signal spacing.

[tradephoric]

Except earlier in the article, the author appears to admit this design is one he created and isn't actually in existence, so there's no real world application and verification it's actually a superior design. Furthermore, note the use of *but it should be efficient*, which means the author has no clue if it would work:

Superstreet interchanges do work and they aren't that uncommon in Michigan (go right to turn left is normal for Michigan drivers).  Here's a few examples of partial superstreet interchanges:

https://www.google.com/maps/@42.9750535,-85.5908314,331m/data=!3m1!1e3

https://www.google.com/maps/@42.7271817,-84.5452441,271m/data=!3m1!1e3