In Toronto there weren't any examples when I last worked there (two years ago). There are apparently some intersections here in the Netherlands which have radar detection which can extend the walk light if they detect someone walking slowly enough that the normal clearence time wouldn't be sufficient. I haven't personally dealt with them, but it seems like a better solution than the braindead Toronto option of just using insanely long Flashing Don't Walk durations which has numerous disadvantages:
- preventing the signal from reacting in meaningful ways to the present traffic demands (including pedestrians)
- increasing the cycle length (and thus pedestrian delays)
- generally making the legal definition of the Flashing Don't Walk (you can't enter the intersection) less and less reasonable, since most pedestrians can actually enter long after the start of FDW and still complete the crossing without causing any issues.
The Netherlands uses a walking speed of 1.2 m/s for pedestrian clearance times, which is the same as Toronto's baseline standard, but Toronto has been moving increasingly to 1.1 m/s or 1.0 m/s on a case-by-case basis as a knee-jerk reaction to complaints. Those complaints are of course genuine problems which can be addressed, but I would prefer intelligent solutions which only create negative impacts on other road users, when there's actually a pedestrian there who benefits from that extra time.
As far as requesting a walk signal, it would be fairly easy to use existing camera detectors to trigger walk signals for pedestrians, but only if there are clearly distinct waiting areas for different directions, and you assume that all pedestrians want to cross the street. Since it would only work in a specific subset of situations, there wouldn't be much point in installing it at all, since pedestrians would still need to maintain the habit of pressing the pedestrian button to call the walk light. The intersections would still need to have pedestrian buttons anyway as a backup system for cases where the camera fails to detect someone.
Yes this is basically my career objective. I've spent the last couple years learning Dutch traffic signal engineering, and I've identified numerous action items we could implement in Canada/USA to make our signals far more responsive to real-time conditions, and much safer for pedestrians and cyclists. I've started assembling them into a series of YouTube videos explaining each concept, but the problem is that after spending 8 hours working on traffic signals, the last thing I want to do when I get home is work more on (a video about) traffic signals. Hence why in the last two years I have only
completed 1 of the 9 planned videos in the series. And also why I've still only completed 1 out of 3 videos about transit signal priority, four years after starting that series. Starting around November I'll be unemployed again so I'm thinking of just taking a few months and pumping out those videos, because I feel it's a huge knowledge gap in North America and youtube is a surprisingly effective tool for educating engineers and planners in Canada.
Signals actually work on the principle of
extending greens, not truncating them. The obvious exception being priority interventions from transit or emergency vehicles, which do indeed truncate them. In general if a light is green but nobody is using it, it's because it has a fixed green time set somewhere and there probably isn't any detector at all for that direction.
If you're interested a technical description of the mechanisms traffic signals use to respond to traffic conditions (even in Toronto!), you could see my blog post
here. You may also be interested in all the videos in my
Traffic Signals playlist.