Babbage

Science and technology

Road safety

The Difference Engine: Safety first

Feb 4th 2011, 16:49 by N.V. | LOS ANGELES

IT IS remarkable how risk-conscious people have become, especially on the road. Sure, some motoring maniacs will always push their luck, causing mayhem for themselves and others—and everyone makes mistakes from time to time, gets distracted, becomes impatient and is, perhaps, not as mindful of other road users as he ought to be. Nevertheless, the statistics for traffic accidents, at least in developed parts of the world, reveal a heartening downward trend.

In the United States, for instance, the latest figures from the National Highway Traffic Safety Administration (NHTSA) show that 33,808 people died on American roads in 2009—the lowest level since 1950. That is still way too many personal tragedies. Even so, it represents a 9.7% decline from the figure in 2008, which was itself 9.7% lower than 2007's. The absolute number of fatalities may grab the headlines, but the more relevant statistic—the fatality rate per 100m vehicle-miles travelled—has also been inching steadily down over the past half century. In 2009, the American rate had fallen to 1.13 deaths per 100m vehicle-miles. Only Britain, Denmark, Japan, The Netherlands and Sweden fared better. For that, traffic authorities everywhere can thank the wholesale introduction of safety-belts and air-bags, as well as tougher drunk-driving laws.

As could be expected, the recession has played its part in reducing the deathly toll on the road—especially among the most vulnerable group, 16- to 24-year-olds. They have suffered most from unemployment and hence have been exposed to fewer hazards on the road. The worry is that there could be a rebound in fatalities once the recovery gets seriously underway and the young resume their reckless driving habits.

While horrifying, traffic accidents are far from being mankind’s greatest scourge. Around the world, they account for 1.2m deaths a year, compared with the 35m people who die from non-communicable illnesses such as cardiovascular disease, cancer and diabetes (5.4m of which are caused by smoking alone). According to the World Health Organisation, some 25m people all told have died in road accidents since horseless carriages took to the streets (the first such fatal accident occurred in London in 1896). That is the same as the number of people who have died over the past 30 years from AIDS.

The irony is that, while the roads are safer than ever, motorists have become more safety conscious. Back in the early 1970s, when your correspondent built a car for himself, he considered its backbone frame—made of pressed-steel sections braced with steel tubing—as state-of-the-art as far as crashworthiness was concerned. With the engine and transmission amidships, the front third of the vehicle was effectively a dedicated crumple zone. Likewise, the rear had strategically placed structural members designed to collapse on impact and mop up excess kinetic energy if shunted from behind. An added virtue was that, being a mere 1,450lb (660kg), the car had very little inertia to overcome relative to most other vehicles on the road, and thus tended to be shovelled down the highway intact when hit from behind (as has happened twice) rather than being crumpled on impact.

Today, though, he considers his beloved 39-year-old car a death trap, and won’t allow his wife or daughter to drive it or ride with him. The reason is not that he thinks it dangerous to drive. Over the decades he has upgraded—on a machine that was inherently safe to start with—the brakes, the tyres and the suspension, and made the frame torsionally even stiffer. As a result, the vehicle now has far more primary safety (the agility, stability and stopping power needed to avoid accidents) than the vast majority of modern cars.

The problem is the vehicle’s secondary safety—the ability to save occupants’ lives if the car is, despite all its primary safety, actually involved in a crash. While the car's original seat belts have been replaced with four-point harnesses, it still has no air-bags, nor any side-intrusion protection. Viewed from the side, its occupants sit within a fragile eggshell of fibreglass. Tee-boned at a crossing, they would be instant spam in a can.

That never used to enter your correspondent’s mind. Nowadays, he thinks about it every time he gets into the car. Put it down to better driver education, more graphic media coverage of road accidents, or simply old age. Yet, the likelihood of his ending his days that way is remote. Statistically, he is more likely to be murdered than to suffer a fatal side-impact.

Without question, the biggest killer stalking the roads today is driver distraction, followed by drink, speeding, fatigue, aggression and the weather. Researchers at Virginia Tech reckon 80% of crashes and 65% of near-crashes involve some form of distraction three seconds prior to the incident.

Despite the fact that it is illegal in many parts of America to use handheld phones while driving, the most common distractions behind the wheel remain texting and dialling. Motorists who text while driving increase their risk of a crash or near-crash 23-fold compared with those who do not. Reaching out for something inside the car represents a nine-fold increase in risk. Dialling causes a six-fold increase (see “Driven to distraction”, October 2nd 2009).

After much cajoling, carmakers are addressing such issues. At a recent get-together between traffic-safety officials and industry experts in Washington, DC, many of the speakers focused on preventing accidents rather than surviving them—in short, a timely re-examination of primary, instead of secondary, safety. Judging from what the NHTSA administrator, David Strickland, had to say, motorists can expect all manner of warning devices to appear inside their cars. One will beep if the driver inadvertently drifts out of lane. Another will screech when a vehicle approaches too fast from the side. Vehicles will talk to one another, warning their oblivious drivers of threats ahead. Rear-view cameras will warn when a child is playing behind a stationary vehicle.

But the technologies NHTSA is putting greatest emphasis on are those that keep intoxicated motorists off the road. Across America, a third of traffic fatalities these days are related to the use of alcohol. One system uses sensors developed by QinetiQ North America, a research and development company that spun out of the British defence establishment, that can measure a person’s blood-alcohol content through the skin. Attached to the steering wheel or the door handle, the device would stop anyone over the limit from driving home. To prevent such sensors from being thwarted by gloves, the vehicle would be activated only if the device received an actual reading of the person’s alcohol level below 0.08%.

Publicly, carmakers embrace such initiatives. Privately, they are leery of them. The added cost is one thing. Customer resistance is another. Then there are the legal liabilities resulting from all the likely false-positive responses that will doubtless lock a proportion of sober drivers out of their vehicles or incapacitate their engines in some way. Lawyers will have a field day.

The security industry has been grappling with similar questions when trying to screen for terrorists among the millions of innocent travellers at airports. The problem is that biometric systems—whether they measure blood alcohol, fingerprint geometries or facial features—do not provide binary yes/no answers like conventional digital systems. By their nature, they generate results that are probabilistic—and hence inherently fallible (see “Dubious Security”, October 1st 2010). Worse, the sensors degrade with age and their data can be corrupted by environmental factors. There are endless ways for a sensor’s accuracy to drift out of true.

Yet there is a real and urgent need for technologies that can keep habitual drunks off the road. According to NHTSA, drivers who were involved in fatal accidents and were over the limit at the time were eight times more likely to have had a prior conviction for drunken driving than drivers involved in crashes who were stone cold sober. Preventing such people from getting behind the wheel might save up to 9,000 American lives a year. Presumably, such savings in life and limb would be similar, or even better, elsewhere.

Readers' comments

The Economist welcomes your views. Please stay on topic and be respectful of other readers. Review our comments policy.

Sort:

Left and Right.

The first fatality was my great great uncle Major Richer on Harrow Hill in 1898(there is a plaque there to mark the event) The driver Edwin Scarth (also killed) was giving a demonstration and when asked what happened when the hand brake and foot brake were both applied he demonstrated causing the centres of the wooden wheels to be ripped out. The Benz dog-cart then overturned killing them both. There were at the time 418 vehicles on the roads of the UK. I think however that three years previously a pedestrian in Lincolnshire was run over and killed. My great Grandmother was the first person to be convicted of going through a red light. She imperiously and indignantly defended herself in court stating that 'she was proceeding along the King's highway and as there were no other vehicles coming and could see not reason why she should stop. ‘Our family's driving has since improved!

Paul Lewis-Berlin

solution to save 250k lives per year?
in-car speed limiter..........ie. you cannot drive faster than the limit.

we went to war for 3k lives over 9-11......but 500k lives per year globally on the roads is just a fact of life.

appicharlak

It is not only sufficient to pay attention to the data gathered through statistics but also it is necessary to look at the underlying process generating the data which is observed prior to committing to investment into more technology of bits and bytes.

It is not true that manifest erroneous behaviours are best examples of human nature and behaviour, as we know, by the facts of accumulated experience that human suffering arises from only three dimensions: from the acts of other persons, beasts and wrath of natural events. Road accidents fall into the category of first kind and therefore, investment into road safety should consider the possibility of unethical trade-offs between private profits and public safety are not made to facilitate investments into unsafe technologies. Accidents are indirect results of bad investment decisions and unsafe technologies are direct causes of them.

Kartellen

In response to the comments about the 35m deaths per year through disease and related illnesses. Whilst yes, they deserve attention, is it not my fault if I smoke or eat unhealthily? Is it my fault if I get hit by a drunken driver? No, it's not and I think it's unfair to suggest that the correspondant is somewhat callous for choosing to feature road safety over something else. By that rationale, very little would ever be reported on as it would be eclipsed by a 'bigger' threat or worry.

canute99

Babbage disappoints. He notes that in the US the fatality rate per 100m vehicle-miles had “been steadily inching down over the past half century.” He goes on “For that, traffic authorities everywhere can thank the wholesale introduction of safety-belts and air-bags, as well as tougher drunk-driving laws.”
The “inching down” has been going on for much longer than the last half century – it began with the Model-T – and long pre-dates seat belts, air bags and drunk-driving laws. In Britain the “inching” rate has been about 5% a year since 1950 and the death rate per 100m vehicle kilometers is now about one twentieth of the 1950 level. The contribution of seat belt and drunk driving laws to this “inching” is statistically undetectable – see “Managing Transport Risks: what works?” (http://john-adams.co.uk/2010/12/02/managing-transport-risks-what-works/).
The inching-down with increasing levels of motorization, noted by Babbage, is a social learning process first described by Reuben Smeed in 1949. It is manifest in countries at all levels of motorization and appears to have little connection with technical or legal safety interventions. Today countries with low levels of car ownership are achieving kill rates per vehicle, with modern imported cars with 100 years of safety technology built into them, as high or higher than were achieved with model-Ts 100 years ago. The inching-own process is now widely labeled, in honour of its discoverer, as the manifestation of Smeed’s Law – try Google.

Glucoseboy

I'm with nirvana. Seems like spending more and more to reach an impossible goal - 0 casualties related to private driving. I'd like to see more money spent on driver education and training rather than self- driving robot cars. Take a cab, a train, a bus, and check all the emails and make all the calls you want. Driving testing exam procedures vary greatly from state to state / province, but everyone ends up licensed to drive on everybody else's roads. More standardization might go a long way.

Anjin-San

It's about high time we should make multiple alcohol/narcotics detectors mandatory in all cars, so that the car will simply not turn unless all the detectors show an "all-clear".

Nirvana-bound

How ironic! By your own admittance, on an average 35m people die worldwide every year from non-communicable illnesses as opposed to 1.2m from auto accidents & yet they get only passing mention, while your primary focus is on improving safety standards for autos! What about the plight of the poor, hapless 35m sick??

How callously indifferent!

KCCM

The next round in road safety (and efficiency) is to take the human driver out of the equation altogether. I'd much rather call up (on my iPhone) a self-drive 'Smart Share' car in town, or have my self-drive Mercedes lock itself into a special lane of the expressway for an out-of-town destination, and catch up on my calls and correspondence (without worrying over that harried parent in the next lane trying to text and negotiate traffic at the same time.)

WT Economist

I live in a city where the majority of those killed or injured in motor vehicle accidents are not in motor vehicles. The are pedestrians or bicycle riders. Are the auto manufacturers considering doing anything about that?

MacMed

Eye tracking devices could alert a driver with progressively louder alerts if the eyes have not looked in the direction of travel for longer than approximately 5 seconds.

Such devices would be bought by safety-conscious drivers. They should not be compulsory until they had evolved to be extremely reliable, and there was solid evidence of efficacy. This was the acceptance route that seat belts followed.

willstewart

The distraction issue is I think insoluble; we just have evolved brains that can be easily distracted. When you consider the early environment, where dangers might appear at any time, you can see why; distraction could save your life!

So an external technical fix sounds good since we have probably already reached the limits of exhortation. Of these by far the most promising is robotic driving; partial at first maybe.

Another possible line is indeed cellular telephony; but we need to be sensible. Headsets are still less common than they should be; can we make them more so? How about phones that can only be used with a wireless headset?

Alice in Wonderland

Jomiku is correct: the unwillingness to be accountable for one's actions necessarily leads one to seek external solutions. That is, solutions that prevent one from making a mistake or that allow one to offload responsibility when errors arise. But this behavior is not restricted to accident prevention on the road; it permeates our society.

The offloading of responsibility is the engine behind big government. The more government bureaucracy takes over, the less room there is for individual action. A certain mindset applauds this usurpaton because it frees people from having to plan and be accountable if their plan goes awry, but it also undermines the ability of those who *can* plan and *want* to plan for their lives.

Ultimately, it is the difference between wanting to be an employee where your actions are directed and responsibility for the outocme of those actions lies elsewhere, and wanting to be an employer who plans, directs, and takes resposibility for the outcome of those plans.

Leaders are few, and followers are many. So when government policy is set for the majority, what you are going to get are policies that are designed for those who want to be employees and seek to offload responsibility to others. That is the nature of a democratic government and thus of a democratic society. Democracry is about freedom of individual action and accountable government, but it is also about freedom from individual accountability and the centralisation of power. And when you look out of your window, don't be surprised if that is what you see.

shaman2

Here in New Zealand we have recently been told that a large majority of accidents are caused by recidivist drunk drivers. Yet they continue to to drive, with no licence until they actually kill someone. We take no action to prevent this, or impose any real punishment. Sounds like the US is the same. How to fix it short of special jails, or compulsory non drinking? The other dangerous group is men between 15 and 25. These two groups (in combination) cause something like 90% of accidents and fatalities.

PSH

Bayes theorem strikes the biometric solution again ("Dubious Security").

However, unlike TSA's shenanigans, we might be able to get a good number on the attempted subversion rate, and therefore a good handle on the effectiveness of the devices.

TSA will no doubt oppose these devices. The truth might really hurt their budget.

jomiku

It's interesting that we always look for a technical solution. As a society, we have trouble punishing drivers because we know that but for the grace of God it might have been us. Maybe we were at a cocktail party after work and got just too tipsy that time. Maybe we got a phone call that we just had to answer while driving.

It's another example of the intersection between individual liberty and the costs that liberty imposes on others. You drink just that once and someone dies and that family bears the cost. We all pay for the cost of caring for those injured.

In most cases, the intersection is merely financial: as in healthcare when Americans want to be able not to buy insurance and then have the rest of society pay for their care. In this case, the intersection is truly personal: my or your fault for having a few too many or for getting distracted by a text message versus the death of someone else. We call those "accidents" because we prefer not to make each of us rigorously responsible for our actions. So we look for technical solutions.

About Babbage

In this blog, our correspondents report on the intersections between science, technology, culture and policy. The blog takes its name from Charles Babbage, a Victorian mathematician and engineer who designed a mechanical computer.

RSS feed

Advertisement

Babbage audio

Trending topics

Read comments on the site's most popular topics

Advertisement

Products & events

Advertisement