Wired.com had a piece this weekend about researchers from MIT who have developed an algorithm to predict vehicles that are likely to run a red light. Previous attempts at predicting this behavior weren't accurate enough to make the technology viable. It appears that researchers have gotten it down to an algorithm with an 85% accuracy.
“Even though your light might be green, it may recommend you not go because there are people behaving badly that you may not be aware of,” said Jonathan How, an aeronautics and astronautics professor who co-created the algorithm.
One of my favorite mantras is that you "can't always arrest you way out of a crime problem". This also goes for traffic problems. It may be that the solution to the danger of red light runners isn't writing more tickets, but instead finding ways for people to avoid dangerous situations in the first place.
Better engineering has significantly improved highway safety whether it's better designed cars or better designed roadways. It also doesn't have the stigma that's associated with traffic enforcement, that is, that traffic enforcement is more about revenue generation and less about making roadways safer.