They are not making them more efficient. They are studying how to make them more efficient. It’s a really big difference because I am not yet ready to place my life in the hands of a CNN-based AI
There was research on this at least 10 years ago where approaching traffic was detected and the lights were dynamically adjusted to improve traffic flow. This is the first time I’ve heard of this since.
Sadly, no one is considering optimising pedestrian traffic flows as they tend to have to wait at beg lights for increasing amounts of time. If we want liveable cities then we need to consider walking as a priority.
It’s technically only illegal here in the US if done unsafely, too. So if there were no cars in sight, walking through a red light is ok. Crossing somewhere other than a crosswalk is also okay. Walking out into busy traffic without even looking is illegal. Also stupid and dangerous.
As I recall it, the system used cameras or maybe radar to detect the cars some distance away so it had time to optimise the lights. The Google system seems to be based on historical data which probably the quickest way to improve using the current technology. My observation is that most traffic lights are set and forget so there are probably huge improvements possible.
As google data is based on mobile phones, they probably do have data on pedestrians as well as cars which could be promising.
To really fix things there will need to be new technology to allow realtime adjustments which introduces the opportunity to detect bikes and pedestrians.
I have concerns about what this would mean for non-drivers. Sometimes you don’t want lights to align for maximum throughput. There are other factors at play here.
Add comment