An item in IEEE Spectrum by Prachi Patel notes the development of a smart traffic system in Pittsburgh. Called Surtrac, the system developed by CMU professor Stephen Smith uses Artificial Intelligence techniques to adapt traffic signals to current conditions.
Prof. Smith's research suggests that Surtrac has reduced trip times 25 percent and idling times by over 40 percent, a significant difference.
Although such a result may obtain in the short term, decreases in congestion may evaporate over time due to the phenomenon of induced demand. In brief, induced demand refers to an increase in consumption of a resource after its supply is increased.
A smart traffic management system that reduces congestion would be equivalent to widening the roadways. That is, reduced congestion would produce more empty pavement at any given time, just as a program of road construction would produce. Historically, road construction programs have proven ineffective in reducing congestion in the long run.
In effect, smart traffic management would fall prey to Jevons' Paradox.
There are more effective solutions, such as congestion pricing or even increased parking fees. However, such schemes are politically touchy since car-loving citizens tend to see them as craven cash grabs. This perception may be what killed former New York Mayor Bloomberg's plan to implement congestion pricing in that city.
The Mayor eventually opted for a smart traffic system, an approach that does not excite the same political problems.
In the final analysis, the popularity of smart traffic management systems may have much to do with their political rather than their engineering features.
In the meantime, you can learn how Swedes in Stockholm have learned to love, or live with, their congestion pricing scheme!