How Much Time Is Spent at Traffic Signals?

transportist.org

10 points·sharkweek··5 comments

I started my business (https://roadometry.com) to solve this problem. In my opinion, the main issue is the lack of good measurement systems for traffic; most cities only monitor a small fraction of their traffic in real-time.

To waste less time at traffic signals, we really need something more like a vision-based system at every controlled intersection. Then timing can be coordinated on a second-by-second level rather than simply re-tuning every few months, or relying on very rough heuristics.

My system currently requires a desktop with a high-end GPU, but I'm looking at different options for porting this to a low-power architecture for a solar installation.

Intersection control with cameras has been around for a while. Depending on the site conditions, it can be less expensive or more accurate to put a camera up on the signal pole looking across the intersection than a sensor loop in the roadway.

Power is not a huge concern, because most intersections are served by utility power. In most places, utility power runs along the roads, and where that's not true, there's often not enough traffic to need a signal.

Telco's sell aggregate data on this.

Do you integrate with SCATS?

The Not Just Bikes Youtube channel has a nice video on Dutch traffic light coordination.

https://www.youtube.com/watch?v=knbVWXzL4-4&t=312s

Reply on news.ycombinator.com