[McLaren Rulez]

Before I begin, this is not a "Is it better spend extra on ABS?" thread. We've had many of those and the answer is obvious. I'm doing a project on the effectiveness of ABS under differing scenarios and need some help. Yesterday, I was surfing through YouTube and came across this video.

YouTube - ABS vs. No ABS

Very interesting video, isn't it? Basically, it is saying that ABS systems are calibrated such that their braking is actually not as effective as threshold braking (i.e. braking at constant force just before the point of lock up). In other words, if, theoretically, a driver knew exactly how much the brake pedal could be pressed before lock up, he will be slowing down faster than an identical car equipped with ABS and the brake pedal fully depressed.

To verify what he's saying, I'd like to get an idea of exactly how ABS works. My basic idea about ABS was that it would not be engaged until one or more of the wheels locked. At this point, ABS would start "pulsing" the brake pedal at several times per second. In effect, that would create a series of micro-lock ups on and off but because any locked up wheel was almost immediately unlocked, the driver still had steering control. What we're seeing is that the ABS equipped car's brake pressure graph would probably be an oscillation; first its too much so the wheel locks, then its too little so it has to apply more pressure if the car is to slow down. The logic behind this video is that a constant braking force is better than the pulsing braking that is done by ABS.

So here's what I want to know.

1) How realistic is it to expect an experienced driver to be able to judge the point of threshold braking such that he beats the ABS system? According to the video, its very much possible but I'd like to get clear hard facts.

2) How far above and below the threshold braking force does the ABS system pulse? For instance, if say the brake pedal had to be engaged at 70% of its maximum force, the ABS will fluctuate its braking force between (70-x)% and (70+x)%. What is the value of x? And does it vary between different cars? The larger x gets, the better the chances of a driver beating the ABS system.

If anyone can get me scientific evidence proving or disproving the claims, I'd really appreciate it. Any info on the working of ABS, microsecond by microsecond, would be really useful too.

Basically, I know its theoretically possible to beat the system when it comes to braking distance and maintaining steering control at the same time, but is practically achieveable? And if ABS does beat the driver all the time, how significantly ahead is an ABS equipped car as opposed to a non ABS car if both are being driven by top class drivers?

Thanks all!

[ImmortalZ]

Unfortunately, ABS calibrations, just like TCS and SC and ECU calibrations, are a closely guarded secret of the automobile industry. It is a very high value intellectual property and nobody is gonna reveal their secrets.

As a rather novice driver, I simply cannot threshold brake in a car. In an emergency, I simply stomp and steer. In a non-ABS equipped vehicle, I crash into the obstacle and die. Simple.

You are asking a lot of subjective questions which will cause a big fight to break out, like it has every time someone asks this very same.

As for the video, I call BS. There are far too many factors involved that a computer can take into account that you can't. Wet patches, pot holes, steering angles... The ABS system takes the wheel into partial lock. It is simply not letting the tyres go into a standstill. It is dipping into the region of partial lock. The video is also showing a simulation of perfect threshold braking. I want to see a driver threshold brake that well in a real life situation.

To tear apart one point: "Steerability is not an option to avoid collision." You don't threshold brake when you don't need to steer. You stomp as hard as you can. In the rain, your car is out of control in an instant. In the gravel, you should be in an EVO . After removing the ABS, "One wheel will lock because your wheels don't have electronics any more". What the hell. I'm confused. You remove the electronics and you complain the electronics doesn't work? Wheels will lock individually if one has better traction. If the discs are work differently. If the tyres are worn differently. If the brake pads are worn differently. Depending on the material of the asphalt. It's our roads - you never know which year the tarmac your left side is resting on compared to the right.

[adnaps]

Originally Posted by McLaren Rulez

So here's what I want to know.

1) How realistic is it to expect an experienced driver to be able to judge the point of threshold braking such that he beats the ABS system? According to the video, its very much possible but I'd like to get clear hard facts.

2) How far above and below the threshold braking force does the ABS system pulse? For instance, if say the brake pedal had to be engaged at 70% of its maximum force, the ABS will fluctuate its braking force between (70-x)% and (70+x)%. What is the value of x? And does it vary between different cars? The larger x gets, the better the chances of a driver beating the ABS system.

Without substantiating with sceintific data I could say that it is impossible for a even a expereinced driver to execute consistent threshold braking in varying conditions of road surface, vehicle speeds, reaction timing etc. Threshold braking in test conditions like in the video is one thing , real life threshold braking under actual emergency conditions is another. ABS should be a definately a more reliable and consistent performer than the human being in real life.

Secondly even in a ABS equipped vehicle, one could execute threshold braking by not crossing the threshold. ABS kicks in only when one of the wheel (or more than one) starts slipping, whereas threshold breaking is expected to be executed before the wheel starts slipping.

Thirdly how close to the threshold (value of x) a driver can go to depends on the drivers skill and has nothing to do with the ABS design.

[McLaren Rulez]

@adnaps
Regarding your first point, I'm trying to get data to prove/disprove that ABS beats the driver. My gut also tells me its more likely that the computer is superior but its proof that I can't find. Regarding your point about not crossing threshold braking, that is perfectly true. But once you pass it, ABS will be active for the rest of the braking period. But in a non-ABS car, you can bring it back down to the threshold point. That's what the video also mentions. And the third bit, regarding the value of x, I was talking about the calibration details of the system. It is a relative comparison between the driver and the system so both the driver's ability and the calibration have to be accounted for to make the conclusion.

@ImmortalZ
I'm not trying to start a fight but I really am doing a project on the effectiveness of ABS. And so I was hoping for some extra information for it. Maybe someone could get data on the calibration of discontinued cars. Or perhaps there are estimates on what sort of calibration values most manufacturers use.

The issue is this. Threshold braking beats pulsing because your deceleration is maximized at the threshold point. In other words, ABS is not as good as the laws of physics allow it to be. The question is whether it is reasonable to expect drivers of skill to outdo it. Even if a driver is not exactly at the threshold, a good driver is going to be quite close to it. Does that make him safer? I realize that the skill of the driver is a big subjective variable that is difficult to account for. Also, it is possible that ABS may be better in some conditions while drivers may be better in others. This is why calibration data would help understand how close ABS is to optimal braking. And I'm sure it is possible to find out how close the average driver gets to optimal braking in a non-ABS car. I also found this study that evaluates ABS and while the general idea is that ABS is good, it also mentions that there was only a small difference in braking distances on dry surfaces. And the scary part in the report was this

All types of run-off-road crashes - rollovers, side impacts with fixed objects and frontal impacts with fixed objects - increased significantly with ABS. Nonfatal run-off-road crashes increased by an estimated 19 percent, and fatal crashes by 28 percent.

Rollovers and side impacts with fixed objects - crashes that typically follow a complete loss of directional control - had the highest increases with ABS. Nonfatal crashes increased by 28 percent, and fatal crashes by 40 percent.

Basically, I am looking at two ideas of interest:

1) Can drivers can beat the computer and how likely they are to do so?

2) Why are some statistics showing ABS equipped cars being more unsafe?

Thanks for the replies so far guys. Please keep them coming.

[dadu]

Originally Posted by McLaren Rulez

Basically, I am looking at two ideas of interest:

1) Can drivers can beat the computer and how likely they are to do so?

2) Why are some statistics showing ABS equipped cars being more unsafe?

Thanks for the replies so far guys. Please keep them coming.

1) No, they can't beat the computer as human reflexes can't match them in panic situations and speeds.

2) Initially, when ABS was introduced, there were more accidents as people assumed ABS to be a quick fix for high speed stopping which was not the case and people were not trained to use their vehicle along with ABS, i.e. brake and steer away. But now thats changing as people get used to it. (this is what I read from the reports published in USA & Australia)

[ImmortalZ]

I believe that the ABS computer is a much better at skirting the laws of physics than any human with their comparatively slow as molasses reflexes and tendency to freeze up in emergencies.

Drivers who can approach and surpass ABS systems are one in a million. They have better things to do like race F1 cars or defy the laws of physics in WRCs. ABS is not for those people. It is for everyone.

[adnaps]

Originally Posted by McLaren Rulez

Basically, I am looking at two ideas of interest:

1) Can drivers can beat the computer and how likely they are to do so?

2) Why are some statistics showing ABS equipped cars being more unsafe?

Thanks for the replies so far guys. Please keep them coming.

@McLaren,


Reg Point 2,

There are two kinds of statistics, the kind you look up and the kind you make up. ~Rex Stout, Death of a Doxy


A statistical analysis, properly conducted, is a delicate dissection of uncertainties, a surgery of suppositions. ~M.J. Moroney

Statistics can be made to prove anything - even the truth. ~Author Unknown

98% of all statistics are made up. ~Author Unknown



Statistics Quotes and Sayings

[MihirC]

Excellent thread McLaren, It was really enlightening to see the video. It actually makes a lot of sense. Thanks a ton for getting this on the forum.

[longhorn]

@ McLaren - Excellent thread.

The advantage with ABS is that since each wheel is independently controlled, the brake force can be adjusted on only the wheel that is about to lock. This is impossible to replicate by a driver as full force is applied equally to all wheels. ABS aids in steering the car in panic situations, where we would normally lose steering control under hard braking. IMO, this is very important, because if we are able to control the direction, we are in a better position to avoid a crash.

I would say ABS is as good as the software that goes into it. Basically how soon you a able to detect a potential wheel lock is most important here. The earlier its done, the more effective ABS will be. The applying and releasing of the brake force to each wheel is done in a fraction of a second, which is impossible to replicate by a human, F1 driver or not. AFAIK, an ABS equipped car is safer than a non ABS equipped one.

[McLaren Rulez]

Thank you for your replies.

I started working on this two days ago. I'm finding a huge amount of data and while the basic premise that ABS is a great safety feature to have is very much unaffected, specific scenarios are highlighting the weaknesses of the system.

I still need to find specific data on whether a driver can beat the computer. So far, we only have opinions that humans cannot match the speed of the computer. No facts. If the computer's pulsing is significantly less effective than threshold braking, it is possible for regular drivers to be as good or better. Conclusive proof is hard to find. If anyone has any idea on calibration of ABS systems, please spill the beans.

At this point, let me introduce another phenomenon. Risk compensation. The basic idea is simple. If you think you are less likely to meet with an accident, you will go faster. Two seperate studies in Munich and Oslo on taxi drivers ended up with a startling conclusion. Taxis with ABS went significantly faster to the point where the benefits of ABS were negated. So both ABS and non ABS taxis ended up having a nearly identical probability of crashing. This is a plausible explanation for the statistics posted earlier. Simply put, we're sweeping all our concerns about meeting with an accident under the ABS rug. Drivers think its invincible: Reality is, its not!

But let's look at the statistics once again. A 30-40% increase in accidents involving a complete loss of directional stability for ABS cars over non ABS cars? That's a seriously frightening figure. Is it completely explained by risk compensation or is there something more?

What if, the ability to steer was a factor? What if you were heading for a crash and ended up swerving at the last moment since your ABS let you? And the swerve was too sudden for your car to handle and therefore, you lost directional stability. And in a panic situation, none of us can really turn the wheel calmly. It will be a sudden jerk. And that's enough to send you spinning, or even rolling over, out of control...

[longhorn]

You can't beat the laws of physics. If you make sudden direction changes at high speed you are more likely to ending up flipping over or sideways - ABS equipped or not. The probablilty increases as the vehicle's centre of gravity increases. ABS is essentially provided to reduce the probability of loss of steering control, not to make you stop faster. I believe people misunderstand it to be something related to help you is stopping faster, which is not the case.

Again, if the nut behind the wheel raises the risk level because his car is ABS equipped, I don't think it's fair to blame ABS for that. ABS, EBD, traction control etc are all safety features. Ultimately, the nut behind the wheel is what matters most.

[SS-Traveller]

Interesting premise on which the videos were created. The author assumes that most roads are smooth and dry, without oil, water, gravel or ice on the surface - which is not too often the case in India. I agree that, given smooth, dry roads surfaced with high-grip tarmac, one could have lesser braking distance without ABS than with - mainly because the coefficient of friction between such a road and a good tyre is pretty close to that between the disc pad and the disc. But that's an ideal world.

Then, in Episode 3 of the series on YouTube, he finally brings up a real-life situation where 2 wheels of a car are off the tarmac and it manages to brake straight. Without ABS, the results would have been catastrophic.

YouTube - ABS vs Non ABS - Episode 3

The author's argument really doesn't cut too much ice - I believe he is trying to prove what a good driver he is, that he can drive better without ABS than with. 95% of drivers (except perhaps those who frequently drive on loose sand or drive competitively) would beg to disagree with him (and that includes me).

As to the option of being able to turn off ABS, I would also be happy to turn off at will things like the ratchet lock of the handbrake, be able to change the F:R brake bias at will through a little knob on the dash, have the speedometer switch between reading digitally and in analogue, in km/h and mph at will, and have the option of operating the wipers with either the left or right stalk. That too is a world not attached to reality.

[aaggoswami]

1) The video is not actual and is an animation/video editing. I have doubts if a non ABS car can stop in slower time if the driver is able to hold on to threshold.

It is now wide known fact that ABS helps in keeping car under control in case of sudden hard braking. I am mentioning hard because this is necessary for high speed braking.

2) In EU, snow on road is not a new thing. Here if there is sudden braking at even reasonable speeds, there are chances of losing control over car. Therefore, ABS is required there and no surprise that this is standard equipment since long there.

ABS actually increases braking distance if the surface is not high friction. This means that braking distance will increase in when the surface is snowy, wet road, sand, dirt on road, etc. But here one can control the car in the event of sudden hard braking.

For dry surface, ABS is good. It actually will help in reducing braking distance.

For me, this ABS vs Non ABS boils down to one thing : SPEED. If you are doing high speed and have to brake suddenly, then ABS helps you to maintain control over your vehicle. At slow speeds, this braking distance maintaining control are not that big an issue.

This debate is endless, and I would like to update myself with some data accumulated from actual testing of cars. An example is G2HC i.e. NHC. The initial version had an issue of wheel lock, but the issue was solved with ABS. Here actually ABS helped. But on the other end, in general, I find brakes of Baleno and even M800 to be adequate. They do the job. And in city where the surface is not that good, often with sand or dirt all over the place, its better IMO to get the wheels locked than to have control and dash with someone.

[KSM-Vtec]

Ya very informative i must say, dint know these vids existed

[mkc15]

Effectiveness of ABS also depends on how many channels (read as sensors) the ABS kit uses. Typically, small cars come with 3 channel ABS; separate channels for front wheels and a common channel for the rear. I'm not sure which cars come with 3 and which ones with 4. Civic comes with 4-channel ABS. I guess ABS+EBD ones are 4-channel kits.