ITE District Six Annual Meeting: Red Light Running Panel July 15, 2002

Camera Enforcement v.
Sound Engineering Practices
A clash of diametrically opposed forces!

Continued - Conclusion

From a Safety Perspective:
Compliance is the Only Best Practice!

Starting with what we know for sure about signalized intersections, an intersection’s safety is determined by the actions of the engineer applying verified best practice that attains compliance, not conjecture, myth, cameras or police officers.

Signalized intersections are not a safety panacea, they reduce the capacity of a roadway, cause traffic delays, often create a bigger problem than they solve, increase air pollution and should not be installed if alternative engineering solutions are available and unless all warrants are met. They should be removed or disabled when they no longer are warranted or could be replaced with less intrusive engineering solutions.

Yellow Interval: The period of time allotted for all motorists (including commercial vehicles) to make a decision to go or not to go and to stop or proceed in a calm and orderly fashion. Codified parameters 3-6 seconds.

All Red: The safety buffer to ensure that all expected traffic that may enter the intersection has cleared prior to giving the conflicting traffic the safe to enter green. This should be minimum of 0.5 seconds and should not exceed 6 seconds. However there are instances where longer all reds have been warranted. On higher speeds roads, grades etc. 2 seconds should be the minimum for through traffic prior to giving the green to conflicting cross traffic.

Compliance and safety is further improved when signals are synchronized, flow or sight distance obstructions are corrected, channelization is improved, lamp size increased for better visibility, green duration is sufficient to clear traffic, and when signal changes are marked in advance when they are not readily apparent because of curves, hills or high approach speeds etc. And the intersection accident history is examined for other possible contributing engineering shortcomings.

In all cases the results need to be plotted, documented and be part of the basis for establishing yellow on the light as well as the red grace camera setting.

Even minor changes in yellow interval timing can significantly increase compliance thereby reducing unsafe entries on red. The cameras have documented that there is no rebound, compliance is maintained and motorists do not abuse increased yellow. (1980 ITE research: same)

These Alexandria and Fairfax County, VA charts clearly show the effects of even small yellow interval changes are dramatic, with no rebound. Similarly one San Diego camera location with a simple 1.5 second timing adjustment, reduced red light citations from 2,262 per month to 205 per month, and this could have been further improved. After correcting the yellow interval, one high violation Vienna VA camera recorded 8 violators in 3 months.

There is a statistical curve of probability around the average time a driver needs to respond to a yellow signal, and the goal is to make sure that virtually ALL of the curve is accommodated before the opposing signal turns green. The kinematic formula is the tool for doing this; it’s only a MODEL of real-world response. It does not CONTROL driver response. If it takes 6 seconds to get all the traffic stopped, including heavy vehicles, buses with standees, old people, distracted drivers, drivers with poor vision, in all lighting conditions, then that is what it takes, and there's nothing anyone can do about it. Branding the violators as sociopath and sending them tickets won't help them get their vehicles stopped.

The ITE kinematic formula, at most should be considered a starting point for establishing yellow duration, not a finish point or to satisfy statutory compliance. One ten wheel dump driver who also trailers a backhoe said it best “the yellows are so short here, I don’t even try to stop anymore.” Traffic control devices that lose the respect of the motoring public are the worst of all, and undermine safety everywhere.

Engineers know that the majority of motorists act in a reasonable and responsible manner and that uniformity in expectations improves safety. Using this thesis a researcher found that a 5.5 second yellow interval default setting provides the best overall compliance. If compliance can be maintained within engineering practices it should be.

If safety were truly the end game here, the answer would be an inverse protocol of procedures for yellow interval timing. The best system wide uniform yellow setting should be a 5.5 second default, except for prevailing speeds in excess of 50 mph where it must be 6 seconds. With a minimum 0.5 second all red grace period, and longer as indicated.

For improved system efficiency, yellow interval timing should only be reduced to whatever level compliance can be maintained. Efficacy test: Zero entries on red for 10 consecutive cycles during peak periods. If the yellow is shortened to a level that causes high entries on red, it's an engineering defect, one that will continue unabated and is document by the cameras. Correct it.


This crisis and negative consequences are to be expected when there is a disregard of safety protocols. The lesson here is that no matter how well meaning and logical they may seem, all actions have unintended consequences. That is why it’s imperative that only field verified practices be used as guidance, and it’s the duty of ALL to make sure the vetting process for standards is based only on findings of fact, and that they are applied nationally with uniform expectation, application and appearance.

The camera debacle has also provided us with an interesting view of human nature because driver behavioral change caused by adjustments in signal timing have been recorded. They’ve documented the fact that entries on red continue virtually unabated despite the cameras; and when the yellow interval is increased by as little as a half second, entries on red have dropped precipitously. Yet we don’t act on it? There is a myriad of whys! None are safety!

Engineering tenets: Laws protect the public by regulating unreasonable or unsafe actions. Actions of a reasonable person should be legal. Most people drive in a safe and reasonable manner. Laws cannot be effectively enforced without the public consent and voluntary compliance. Moreover, safety research has found the safest practice is to set all traffic control devices to guide traffic to accommodate what this publics’ consensus has found to be safe and reasonable. Signal timing is no exception.

If there are large numbers of violations or accidents, try another prescribed countermeasure until they are all but eliminated. In signal timing, this may mean only a small incremental increase in yellow duration or a synchronization of the signals or an increase in the all red clearance cushion or new signal heads that are distinct from the background or an increase in lamp size or new delineation or pre warning signs or all. What we know for sure is that the cameras do not solve any of these problems and in most cases they have actually increased accident rates.

There is no middle ground here. If you’re an engineer and safety is your concern, sound engineering practice properly applied can and do reduce accident rates. The tools are there. Yellow interval timing is one of them. Use it. How much? Whatever it takes to accommodate the traffics needs!

As to the rest, if you choose to maintain inadequate yellow clearance intervals, that's your choice but you surely can’t blame motorists for your actions. It’s your duty to correct defects, and you’re personally liable for unsafe actions and the cameras are making the record. This fact will not be lost on attorneys either. Because of the shear number of violations that all viable enforcement camera systems demand for sustenance, it's a Scarlet Letter of failed engineering.

For safety, fair laws, less pollution and more efficiency out of our roadways, engineering will always be the best solution for all. There are no substitutes or short cuts.

Power Point Presentation   RLR Timeline   1  2  3  4  5