Leave a buffer for your bumper: Study contradicts practice of traffic light tailgating

November 27, 2017, Virginia Tech
Jonathan Boreyko (left) and graduate student Farzad Ahmadi monitor pedestrian spacing as a line empties at The Cube in Virginia Tech's Moss Arts Center. The team found that people move slowly but accelerate quickly, meaning packing tightly in lines increases pedestrians' chances of moving through faster. Credit: Virginia Tech

When pulling up to a traffic light, most drivers get pretty close to the car in front of them, leaving just several feet of space between their bumper and the next.

The practice of packing tightly at is widely accepted. Traditional thinking says the closer a car is to a traffic light, the more likely that car will be to pass through the intersection before the light turns red again.

Thanks to new research by Virginia Tech College of Engineering professors and students, now have a good reason to dismiss this faulty line of roadway intuition.

The study, published this month in the New Journal of Physics, used video cameras attached to drone helicopters to capture footage of cars accelerating through a traffic light on the Virginia Tech Transportation Institute's Smart Road. By systematically controlling the packing density of the cars, the researchers discovered that any decrease in distance to the light was completely offset by the time it took for cars to regain a comfortable spacing before drivers could accelerate.

Drivers who pack tightly at intersections do not increase their chances of making it through the light, and tailgating at traffic lights can also lead to more rear-end collisions.

"We varied the bumper-to-bumper spacing between cars by a factor of 20 and saw virtually no change in how much time it took for the cars to pass through the intersection when the light turned green," said Jonathan Boreyko, assistant professor in the Department of Biomedical Engineering and Mechanics. "The results mean there's no point in getting closer to the car in front of you when traffic comes to a stop," he said.

Drone footage revealed that cars take more time to accelerate when they stop closer together than when they stop farther apart. The researchers conceptualized this aspect of the Smart Road experiment as the latent heat of transitioning from a solid, or stopped, phase to a liquid, or moving, phase. Credit: Virginia Tech

The inspiration for the research first came to Boreyko when he was sitting in traffic one day. Noticing that cars had to wait for the car in front of them to regain a safe spacing before they could start moving again, he hypothesized that, contrary to popular opinion, it might actually be better for cars to stop farther apart from each other when idling at a traffic light.

He teamed up with Farzad Ahmadi, a fourth-year Ph.D. student in Virginia Tech's engineering mechanics program and the study's lead author, to investigate.

Using 10 volunteer drivers in identical vehicles, the researchers staged a series of experiments at the traffic light on the Virginia Tech Transportation Institute's Smart Road. Drivers systematically lined up at the light in a set of distances ranging from 1.25 to 50 feet, and a drone helicopter hovering overhead captured controlled bird's-eye-view footage of the traffic as drivers accelerated through the light.

Analysis showed that the time required for all cars to pass through remained relatively fixed, give or take about one second, for spacing distances up to 25 feet.

The two researchers used the thermodynamic concept of latent heat, the energy that a system loses during melting or evaporation, to describe what happens to cars stopped at a traffic light. Vehicles are jammed into a "solid phase" at a light and must waste energy "melting" back into a "liquid phase" before they can actually move through the intersection.

Boreyko and Ahmadi wondered if latent heat would have such a dramatic effect on other systems, such as slow-moving pedestrian traffic. Should people waiting in lines space themselves closer together or farther apart in order to move through more quickly?

In video footage captured by a drone helicopter, volunteer drivers (left) approach a traffic light on the Virginia Tech Transportation Institute Smart Road as part of an experiment on optimal car spacing. Researchers in the Department of Biomedical Engineering and Mechanics discovered that packing tightly at traffic lights does not increase a driver's chances of being able to pass through the light before it turns red again. Credit: Virginia Tech

The researchers set up a second round of experiments in The Cube at Virginia Tech's Moss Arts Center, a highly adaptable theater and laboratory equipped with synchronized cameras. Undergraduate students Hunter Morgan, Josam Waterman, Pat Greer, and Will Doty - all in the engineering science and mechanics program and co-authors of the study - added a few conditions to their senior design experiments on human crowds to test Boreyko and Ahmadi's hypothesis.

"Latent heat had almost no effect for a line of pedestrians," said Boreyko. "The closer people got to each other, the faster they could empty the line. We realized that people move very slowly, but can accelerate very quickly, which minimizes the lag effect we saw with the cars at the traffic light."

The study's findings suggest that both pedestrians and drivers alike could see considerable benefits when taking a mindful approach to packing density in lines.

"Pedestrians waiting in a line should get as close to each other as possible if it's important for the line to empty quickly," said Boreyko. "But when you encounter a traffic jam or stop at a light, keep a safe and comfortable distance. You can just maintain whatever spacing you had when you were driving at full speed. You won't lose any time, but you'll reduce the odds of an accidental rear-end collision."

Ahmadi agreed with Boryeko's conclusion.

"When my father was teaching me how to drive, he told me that to prevent an accident, you should stop so you can easily see the rear bumper of the car in front of you at a ," said Ahmadi. "I've never done that until I analyzed the data of this experiment."

Explore further: Traffic signal countdown timers lead to improved driver responses

More information: S Farzad Ahmadi et al. Latent heat of traffic moving from rest, New Journal of Physics (2017). DOI: 10.1088/1367-2630/aa95f0

Related Stories

Making driverless cars safe for pedestrians

August 18, 2017

Right now, there are two ways to be safe crossing a road: Wait until no cars are close by, so there's enough time to make it to the other side of the street – or communicate with oncoming drivers. As the number of pedestrian ...

Hit all green lights with new Audi gadget

October 24, 2008

(PhysOrg.com) -- If you drive down the same roads every day, you probably know the patterns of familiar stop lights - how long they take to turn green, or when the green arrow will appear. For light-conscious drivers or anyone ...

Recommended for you

First-ever colour X-ray on a human

July 12, 2018

New Zealand scientists have performed the first-ever 3-D, colour X-ray on a human, using a technique that promises to improve the field of medical diagnostics, said Europe's CERN physics lab which contributed imaging technology.

30 comments

Adjust slider to filter visible comments by rank

Display comments: newest first

michbaskett
3.8 / 5 (8) Nov 27, 2017
Encouraging people to keep an even larger distance between them and the car in front of them ignores the reality of dense traffic during rush hour. I can see making getting from here to there take twice as long using tactics such as this. If I can't get between light A and B because I am prevented from getting past light A it's irrevelvant. I'm still stuck and getting from here to there takes longer.

It also ignores the people who are oblivious to the outside world because they have some electronic piece of junk to which they give all their attention. The distance between cars makes no difference in that regard. The can't hear the world outside honking and screaming at them to get a move on.

This might work for places where the level of traffic is always light, but in any large city it's of no value.
overinvested
4 / 5 (6) Nov 27, 2017
Try doing the report again with random reaction time and acceleration. This research looks too controlled as michbaskett comments.
pntaylor
4 / 5 (6) Nov 27, 2017
"When my father was teaching me how to drive, he told me that to prevent an accident, you should stop so you can easily see the rear bumper of the car in front of you at a traffic light,"

The man who taught me to drive, 45 years ago, told me to stop where I could see the rear tires
of the car front of me touching the ground and I always have, because it is a wise thing to do.

You cannot control what others do but you can control what you do. If you get too close to the car in front of you and someone hits you in the rear, pushing you into the car in front of you, you are responsible for the damage to the car you hit.

The farther you are from the car in front of you, the harder someone has to hit your rear, to cause you to damage the car in front of you.
Merrit
1.7 / 5 (3) Nov 27, 2017
Shorter distance between the cars going through the intersection is the important part for getting more cars through.

When self driving cars is perfected it can both reduce accidents and traffic
fishnuke
4.7 / 5 (7) Nov 27, 2017
Absolutely terrible advice. Cars are not bunching at lights to try to "squeeze through" as this over-simplified and over-controlled study assumes. They are bunching to prevent blocking other nearby intersections, and in particular the turn lanes of the immediate light. Both common situations completely lacking here.

Ever since this idiotic trend to leave a car length or more between vehicles started, I have seen so many people missing the light since they can't get into their turn lane. While the self-righteous space holder sits there oblivious in their bubble, obstructing traffic, while feeling really good about themselves.

This is excusing extremely selfish behavior, which impedes the flow of traffic by promoting early gridlock spread to other intersections and denying other people access to the turn lanes.

Redo with fixed cameras at lights at several adjacent intersections, and count how many cars get needlessly blocked by this rude behavior.
dogbert
4.3 / 5 (4) Nov 27, 2017
The study also ignores the predictable effect of leaving a large distance between your car and the one in front of you. In heavy traffic, cars from other lanes will fill the gap. Even during driving, a large gap encourages other cars to cut in front of you. If you continue to leave a wide gap between your car and the car in front of you, your progress will slow to a crawl as every time you back off, someone else will move in front of you, causing you to back off again.

People tend to maximize their ability to progress toward their destination. Artificial changes to their choices generally hinder rather than help.
Drjsa_oba
4.8 / 5 (5) Nov 27, 2017
I read this study and was immediately driven to comment here. However, it appears that all the people commenting have had the same issues with the study as I have.
1 Leaving gaps makes the queue longer, although it has little effect on the individual actually getting through the light, a longer queue may make traffic congestion further from the light.
2 Leaving a gap encourages others with less good sense to try and move into the gap from other lines; they will estimate that your line will be faster than their line because it has less cars in it.

Human behaviour being what it is, the compressed queue is not necessarily wholly motivated by the desire to get through the lights sooner. But I admit there is a factor.
clint420601
1.5 / 5 (2) Nov 27, 2017
Regarding the spacing in the queue. The safest distance is that which just kisses the bumper of the car ahead of you. That way, should a driver coming up behind you fail to see the stopped vehicles, the Force of the impact will be spread over a greater Mass resulting in a lower Acceleration [f=ma]. Additionally, the crumple zone in the rear of the vehicle ahead of you will help your vehicle's front and rear crumple zones absorb some of the energy. Two of my colleagues were involved in such an accident and were severely injured. I suspect they would have fared better if there had been several vehicles ahead of them to help absorb the energy.
Nik_2213
4 / 5 (1) Nov 27, 2017
If you leave a modest gap, you are not only protected against being shunted into car ahead, with all the insurance squabbles about non-mitigation, you may *anticipate* and move in convoy with car in front, thus getting through *quicker*.

Tailgating at lights is related fallacy to nosing over stop-line. If you leave a few feet, you may anticipate gaps and be rolling sooner than the nosed-out ijit beside you, who must leave rubber.

One exception is when queue backs up in a slip-lane to block the main traffic. Then, every inch counts. Of course, it would really, really help if people used their turn signals so you knew which way they wanted to go...
yep
5 / 5 (2) Nov 27, 2017
Cars that leave huge gaps reduce the amount of traffic that can get in the turn lane on an over pass as well as block turn lanes reducing flow in general as stated above. I will remain skeptical of this and would rather see real world data of cars at intersections not involved in a test study. In the real world cars leaving two car lengths cause road rage!
ddaye
4.5 / 5 (2) Nov 27, 2017
We often confuse what's possible with what's statistically likely in average populations.

I stop close to the car ahead at a light, but unlike most drivers, I occasionally begin accelerating at the same time as the driver ahead. Just a bit less acceleration so that the gap increases gradually to a normal one at normal speeds. Virtually every time there's a much larger gap behind me than ahead.

If everyone did this, there'd be significantly more cars getting through lights after a stop, but it requires a level of attention, a type of attention, and reaction times that would be very unsafe to teach or expect of the general population. But self driving cars in large numbers would have the uniformly high capabilities to make a big difference.
carbon_unit
5 / 5 (4) Nov 27, 2017
As others have pointed out, large separation between stopped cars will likely cut off turn lanes, impeding flow. It also sometimes goes the other way. Widely spaced cars in the turn lane can back up into the through lanes producing an even worse impediment.
poksnee
5 / 5 (1) Nov 28, 2017
What is the point of the buffer?
The farther you have to travel to get through the intersection the longer it takes. It does not matter if the distance to the intersection is filled with cars or empty space.
antialias_physorg
5 / 5 (1) Nov 28, 2017
What is the point of the buffer?

I think this somewhat depends on the cars you have in front of you (this from my experience as someone who drives a car that's very close to the road).
If you have visibility to the next few cars in front of you then leaving more space is advantageous because you can already start rolling in sync with cars further ahead (i.e. the entire conga-line starts moving at once).

If you have something that blocks your line of sight (SUV/truck), then closing the gap may be advantageous because you can only start accelerating once you see the car immediately in front of you move (which causes a wave propagation instead of simultaneous acceleration).
antialias_physorg
1 / 5 (1) Nov 28, 2017
Let's do the math here:
Example: 10 cars. Each car has length c and distance between cars is x

Case 1: long distance ( x_long ) between cars and perfect visibility.
All cars can start moving at once so you get the 10th car which needs to travel a distance of
s = 9c + 9x_long
at constant acceleration a the time to reach the traffic lights is
t_case1 = sqrt (2s / a) = sqrt ((9c + 9x_long)/a) = 3 sqrt((c + x_long)/a)

Case 2: no visibility or distance so short that immediate movement is too 'optimistic'. I.e. cars will start as soon as they see the car before them moving (after a reaction time r). Distance between cars is x_short.
s = 9c + 9x_short
t_case2 = 9r + ((9c + 9x_long)/a) = 9r + 3 sqrt((c + x_short)/a)

So as long as reaction time is greater than
(sqrt(c+x_long)-sqrt(c+x_short))/(3sqrt(a))
it's better to have x_long than x_short.


Pratyeka
5 / 5 (1) Nov 28, 2017
In formula 1 racing, as well as other kind of racing, all the car start at the same time when the signal is given. They don't wait for the car in front of them to move away before accelerating.
Unfortunately, very few people have the mental capacity to focus on simple driving techniques.
antialias_physorg
5 / 5 (1) Nov 28, 2017
In formula 1 racing, as well as other kind of racing, all the car start at the same time when the signal is given. They don't wait for the car in front of them to move away before accelerating.

Sure, because they trust the guy in front implicitly to be on guard and get moving immediately (after all: it's their job to have the best possible reaction time). Formula 1 cars are also very close in terms of their acceleration capabilities which cannot be said for a mixed line of cars/SUVs/pickups/trucks/motorcycles/whathaveyou.

Note also that formula 1 races start with crashes a lot more often than your average traffic light line, so I wouldn't advise this kind of behavior for everyday use.
Ironwood
5 / 5 (1) Nov 28, 2017
"When my father was teaching me how to drive, he told me that to prevent an accident, you should stop so you can easily see the rear bumper of the car in front of you at a traffic light,"

The man who taught me to drive, 45 years ago, told me to stop where I could see the rear tires
of the car front of me touching the ground and I always have, because it is a wise thing to do.

You cannot control what others do but you can control what you do. If you get too close to the car in front of you and someone hits you in the rear, pushing you into the car in front of you, you are responsible for the damage to the car you hit.

The farther you are from the car in front of you, the harder someone has to hit your rear, to cause you to damage the car in front of you.


That isn't true. If someone hits you in the rear and pushes you into the car in front, you are NOT responsible.
Merrit
not rated yet Nov 28, 2017
Also, the point of being able to see the tires of the car in front is so that you can go around them if necessary. Everyone seems to be forgetting this point.
Shakescene21
not rated yet Nov 28, 2017
Ironwood is correct. My wife was in a 4-car collision, in which she rear-ended a car and pushed it into the next car forward. Our insurance company was responsible for damage to both cars.
Shakescene21
not rated yet Nov 28, 2017
This article is irrelevant to driving in dense urban traffic. These experiments were conducted near the Virginia Tech campus, which is a in a fairly low population density area in Southwestern Virginia. The drone photo shows conditions that are totally unlike the crowded urban traffic in the DC suburbs of Northern Virginia.
antialias_physorg
5 / 5 (2) Nov 28, 2017
Also, the point of being able to see the tires of the car in front is so that you can go around them if necessary. Everyone seems to be forgetting this point.

I have a much simpler maxim: Always drive under the assumption that everyone else is a total idiot who could hit a wall/slam on the brakes at any time.
If you do this then you will leave enough space so you will never get into a hazardous situation.
hb_
not rated yet Nov 29, 2017
@fishnuke. You are quite right. Long distances make long ques, which in turn blocks other lanes. It is a terrible advice that the authors give the readers.
hb_
not rated yet Nov 29, 2017
The modelling of the driver behavior does not include variations of distance car-to-car when driving. I frequently find myself blocked from a green light when the car in front of me chooses to maintain, say, 10 car lengths to the next car. Also, I find it baffling that drivers are so ... stupid ...

You see, when I see the light turning, I start to roll slowly even if I am 100 yards from the intersection. I watch (!) the traffic light even if I am not the front runner, but most drivers seem to watch only the bumper in front of them. "Oh my, bumper in front moving, time to engage the motor" and the next driver "Oh my, bumper in front moving, time to start engage the motor" a.s.o. As a result, the queue stretches out when the light is green to a fantastical length...
hb_
not rated yet Nov 29, 2017
Let's do the math here:

....

Case 2: no visibility or distance so short that immediate movement is too 'optimistic'. I.e. cars will start as soon as they see the car before them moving (after a reaction time r). Distance between cars is x_short.
s = 9c + 9x_short
t_case2 = 9r + ((9c + 9x_long)/a) = 9r + 3 sqrt((c + x_short)/a)

...



@antialias_physorg; Why do you assume that short distances equal no visibility? And accelerating when seeing the traffic light turn, or when seeing 5 cars ahead accelerating rather than when seeing the bumper in front moving is primarily a function of intelligence, not of visibility. Your model is flawed and your calculations thus show nothing.
antialias_physorg
not rated yet Nov 29, 2017
Why do you assume that short distances equal no visibility?

Actually I do not assume this in my example (notice the 'or' in "no visibility or distance so short that immediate movement is too 'optimistic'") . Visibility does not change the situation for short distances, because you still have to wait till the guy in front of you moves plus the reaction time - even with perfect visibility.

Imagine the extreme case (you're right up to the bumper of the car in front of you). If everyone hit the accelerator 'on trust' that everyone in front of you moves at the same acceleration rate and with the same reaction time this would cause a crash at every traffic light.
tallenglish
not rated yet Nov 29, 2017
This kind of makes sense, if cars are close together you will get the caterpillar effect where people are accelerating and breaking constantly - slowing everything up. If you leave the space there is no breaking and everything is just smoother.
Captain Stumpy
not rated yet Nov 29, 2017
@A_P
I have a much simpler maxim: Always drive under the assumption that everyone else is a total idiot who could hit a wall/slam on the brakes at any time.
If you do this then you will leave enough space so you will never get into a hazardous situation.
this actually has a name - it's called Defensive Driving

Captain Stumpy
not rated yet Nov 29, 2017
@Ironwood
If someone hits you in the rear and pushes you into the car in front, you are NOT responsible.
yes and no

this entirely depends upon how "close" you are to the vehicle and how fast the other vehicle which impacted your rear is travelling

if you can't see the tires of the vehicle in front of you resting on the pavement/road, then the officer who investigates has the option of citing you for being too close

the officer has a means to plug in vehicle data (found on the door placard - curb weight, etc) into a formula to establish force, speed, etc, which then determines your distance from the secondary collision - states usually have a minimum distance required for stationary situations (at lights, traffic, etc) listed in the DMV training manual

though some states allow the responding officer leeway in this decision, it can be overridden by the accident investigator upon further evidence

TheGhostofOtto1923
not rated yet Dec 03, 2017
You cannot control what others do but you can control what you do
Brake checking is an effective way of controlling what others do. I find that even speeding up and then slowing down repeatedly can make tailgaters back off.

This works especially well during acceleration after a stop when pinheads are right on your tail. Abruptly letting up on the gas backs them right off.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.