Green Lights Too Short When Lanes Unevenly Loaded

[Brian556]

I've noticed that when lanes are unevenly loaded, for example, there are two cars in the left lane, but 12 in the right lane, the green light will be way shorter than usual, screwing the people in the lane with more vehicles.

This is apparently a built in feature of signal controllers, but it needs to go.

There is a place where the right lane is closed for utility work daily 9-4 appx.
Today, I noticed the signal give a very short green for the direction of traffic with the closed lane. The green time was way shorter than it would be under normal circumstances.

This is causing traffic to back up worse than it should.

[NE2]

This isn't the state/local DOT.

[Brian556]

The one with the lane closure is FM 407 at Morriss Rd, located in and maintained by Flower Mound.
The other is I-35E NB at Corinth Pkwy in Corinth, maintained by City of Lewisville by contract w/TxDOT. The lanes unevenly load because a lot of the traffic turns right at the next street after the signal; headed to NCTC.

[jeffandnicole]

Could be something as simple as a bad traffic detector. Have you tried reaching out to their traffic light/maintenance dept and have them take a look at it?

[roadfro]

I've noticed that when lanes are unevenly loaded, for example, there are two cars in the left lane, but 12 in the right lane, the green light will be way shorter than usual, screwing the people in the lane with more vehicles.

This is apparently a built in feature of signal controllers, but it needs to go.

There is a place where the right lane is closed for utility work daily 9-4 appx.
Today, I noticed the signal give a very short green for the direction of traffic with the closed lane. The green time was way shorter than it would be under normal circumstances.

This is causing traffic to back up worse than it should.

Signal controllers don't usually shorten the green for unbalanced lane flows as a "built-in feature".

I'll assume the signal is using detector loops embedded in the pavement. Most of the time, the loops in adjacent lanes are linked together so that a vehicle in either lane sends an identical call to the controller. I.e. the signal controller receives a detection call so it knows there is a vehicle on the approach, but it doesn't  know whether that vehicle is in the left lane or the right lane.

A likely scenario in this situation is that a loop detector signal got cut during the utility work. When a detection feed is lost, the signal controller usually registers a "constant call" for the phase and will default the green time to whatever backup setting has been stored for that phase (since it can't detect demand in real time) i.e. either run the minimum green, run the maximum green or run a preset default green length. I would bet that the signal is running the minimum green for this approach each cycle, which is probably much shorter than the typical green phase received based on detection.

[Brian556]

Quote from roadfro;

Signal controllers don't usually shorten the green for unbalanced lane flows as a "built-in feature".

I'll assume the signal is using detector loops embedded in the pavement. Most of the time, the loops in adjacent lanes are linked together so that a vehicle in either lane sends an identical call to the controller. I.e. the signal controller receives a detection call so it knows there is a vehicle on the approach, but it doesn't  know whether that vehicle is in the left lane or the right lane.

A likely scenario in this situation is that a loop detector signal got cut during the utility work. When a detection feed is lost, the signal controller usually registers a "constant call" for the phase and will default the green time to whatever backup setting has been stored for that phase (since it can't detect demand in real time) i.e. either run the minimum green, run the maximum green or run a preset default green length. I would bet that the signal is running the minimum green for this approach each cycle, which is probably much shorter than the typical green phase received based on detection.

Both have image detectors. Neither consistently does this. I've noticed that whenever a signal does this, the next green is typically full length.

The one in Corinth just got put on flashing mode because the overpass next to it was torn down.