The future of the Traffic signal

[mrsman]

I agree with this statement.  You can set detectors not to respond immediately!  On Hylan Boulevard on Staten Island and on the Grand Concourse, modern signal technology would help improve the steady flow.

Back to Manhattan, Central Park West has many signals that are on one way streets inward that are only there for pedestrians as the park is to the opposite side of the street.  How much unnecessary stops does one make for nobody even crossing the street vehicle or pedestrian?  You could advance one more block if you have the buttons installed where the pedestrian activates the signal and also it can be delayed till the next cycle for the light on CPW to turn red keeping sync with the rest of the signals.

Like someone once said on this forum how behind in the times NYC is with technology as he pointed out it was rather recently NYC got fiber optic cables installed.  The same with traffic signals I can see and I believe he said the same thing when he posted that.

BTW, those side mount signals on Fifth Avenue look horrible.  Mast arms would improve the look around Rockefeller Center tremendously!  Plus some intersections like at 42nd Street have them already put in place as well as a few other intersections rather recently too.

Washington DC is pretty similar to New York in that the vast majority of the signals are on pure timer.  They don't even have a push button for the pedestrians.

But one place where they do tend to have push buttons are mid-block pedestrian crosswalks or on similar intersections where a one-way street at a T intersection goes in the direction away from the T.  (Like 3rd and Madison near the Capitol) Since there are no cars on the minor direction, the ped should push the button to cross the street.  In fact, there are usually relatively large signs at such places reminding peds to push the button at these corners, since it's not required at the vast majority of corners.

[tradephoric]

In dense central business districts, pre-timed signals work quite well.  There are 2,800 traffic signals in Manhattan alone.  The cost involved to install and maintain such a large detection system doesn't seem justified.

Keep this in mind... No detection can be better than bad detection.  Do you really trust that NYC DOT would be capable of providing reliable detection for all 2,800 signals?  At any given time, 10% of the detection would be broken/faulty leading to potentially worse traffic flow in Manhattan than what a pre-timed system could provide. 

[roadfro]

Tradephoric does have a point though. Pre-timed signals do work especially well in a dense urban core, especially if the streets are laid out on a grid and/or if the streets are one way. (My understanding of New York City--I've never been there personally--is that there is a pretty solid two-way grid system on most streets but not a lot of left turn phasing, so at least some decent progression should be possible.)

I remember being especially impressed by the street grid in downtown Portland, OR when I was there several years ago. One way streets and pre-timed signals were great. They worked in such a way that you always hit a green wave if traveling right about at speed limit. The street spacing and signal timing was also such that if you happened to be walking at a steady pace against the flow of vehicular traffic, you nearly always hit the walk signal at the intersections.


With that said, side street detection can be beneficial to arterial coordination in the absence of ideal conditions such as an evenly spaced grid.

[mrsman]

Tradephoric does have a point though. Pre-timed signals do work especially well in a dense urban core, especially if the streets are laid out on a grid and/or if the streets are one way. (My understanding of New York City--I've never been there personally--is that there is a pretty solid two-way grid system on most streets but not a lot of left turn phasing, so at least some decent progression should be possible.)

NYC is made up of 5 different boroughs.  Manhattan is largely a one-way grid.  The island is much longer N/S than E/W.  So for most of Manhattan, you have about 8-15 avenues going N/S, the vast majority of them are one-way and relatively wide.  Traffic favors the avenues and the signal progression favors the one-way avenues.  You also have many E/W streets, about 1/20 of a mile apart, but most of them are one-way and only allow one lane of traffic.  The street progression works great in Manhattan.

Staten Island has no significant regular grid.

The Bronx loosely continues Manhattan's grid, but we don't have as many avenues as tightly together as you do in Manhattan.  Many of the avenues and streets aren't significantly orthogonal and most major avenues and streets are two-way instead of one-way.

Brooklyn and Queens have several grids within their layout that sometimes meet at odd angles.  (I.E. see Dahill Avenue in Borough Park where the Brooklyn grid of numbered avenues and streets meets the Flatbush grid of East (and West) numbered streets and lettered avenues, Broadway also cuts across two grid patterns).  Unlike Manhattan, the largest avenues and streets are largely two-way instead of one-way.  Progression does not work as easily as it does in Manhattan.  But at times, since there is a regular grid, the traffic controllers are programmed to provide priority for one of the directions (like towards Manhattan in AM rush, or away from the baseball stadium on Northern Avenue when the baseball game is supposed to finish, etc).  Of course, traffic progression in such a case is miserable if you're going the other way.

[realjd]

I'm of the opinion that they'll remain largely unchanged visually but will become smarter on the back end. Once the signals can communicate with each other or with a central traffic system, computers can automatically adjust the timing to suit current traffic conditions.

Once driverless cars become more common and car-to-car communication standards develop, I don't doubt that there will at least be some research done on tying the automated cars in to the traffic control network. This would allow the traffic planning computer to assign suggested speeds to the cars to minimize delays through the whole network. For example, if a traffic signal could see incoming groups of cars and calculate when they'll arrive at the signal, they could then suggest that the cars from the north slow down and the cars from the east speed up so they both can reach the signal on green and not have to stop.

[roadfro]

I'm of the opinion that they'll remain largely unchanged visually but will become smarter on the back end. Once the signals can communicate with each other or with a central traffic system, computers can automatically adjust the timing to suit current traffic conditions.

What you describe is referred to as an adaptive signal system. While not completely unknown, this technology is not very widespread -- primarily due primarily to cost, and less so for infrastructure needed (comm systems, additional detection, etc.). In many places, communications between signals and a central hub simply allows for automatic clock adjustments and remote manual upload/changes of signal timing patterns (as opposed to driving to the signal to load new timing plans).