Helmet distractions

ghost bike utah

A “ghost bike” marks an intersection where a bicyclist was killed in Utah/Wikimedia Commons

The National Transportation Safety Board (NTSB) on Tuesday issued its analysis of why increasing numbers of bicyclists are dying on U.S. roads even though some – especially women – now don’t ride because of the risks or say they are riding less because of the danger.

The concrete recommendations in the report seem almost sure to get lost in a brouhaha over helmets.

“Today’s report focuses on how we can prevent collisions between motor vehicles and bicycles by creating an infrastructure that separates bicycles from other traffic because preventing a crash from happening in the first place is our top priority,” NTSB Chairman Robert Sumwalt said when he introduced the agency’s analysis.

That might have been the focus of Sumwalt and the NTSB board, and it is the reason the majority of cyclists die. But what attracted most public attention was the very last recommendation in the report: the consideration helmet use be made mandatory.

The media quickly hit that hot button and tagged Sumwalt’s focus on preventing collisions as among “other” safety ideas.

“The National Transportation Safety Board this week said laws mandating helmet use, among other safety actions, would reduce traffic fatalities involving cyclists but the recommendation is getting pushback from bike advocates” was the lead on the Washington Post story.

There is no argument that bicycle helmets will in some crashes save lives. There is some debate as to their affects at a population level. There is evidence that helmets influence the behavior of people wearing them, and that because of this cyclists wearing helmets take more risks than those who are helmetless.

And there are a variety of studies that document that cycling, or many other forms of aerobic exercise, help people live longer and healthier lives. All of those things factor into whether the relatively few lives saved by bicycle helmets would make a difference at a population level.

Helmet realities

But helmets can and do save lives, and they could save more.

Data from the Centers for Disease Control (CDC) indicate that requiring passengers in motor vehicles to wear helmets would likely save thousands of lives. More than 10,500 of the 56,800 people who died of traumatic brain injuries (TBIs) in 2014 were motor vehicle occupants, the CDC reported. 

Not all of those people would have lived if they were wearing a helmet just as helmets will not save all cyclists from deadly TBIs. But given that motor vehicle helmets are far more substantial than bicycle helmets, the ratios of survivors would likely be greater.

If even a tenth of motorists survived thanks to a mandatory, motor-vehicle helmet law, the number of survivors would actually outnumber the 857 bicyclists killed by motor vehicles in 2018. The NTSB offered no estimate of how many of those cyclists might have lived if they had been wearing helmets, but surely some would have.

The same could be said of some of the 6,227 pedestrians hit and killed by cars in 2018 – the highest number since 1990, according to the Governors Highway Safety Association.

Most U.S. motorcycle drivers and passengers – 71 percent, according to the International Safety Council – now wear helmets. But like the other “vulnerable road users,” as Sumwalt described them, their odds of surviving a crash involving a car or truck are not good even with a helmet.

Almost 5,000 motorcyclists or their passengers died in 2018, according to the National Highway Traffic Safety Administration (NHTSA). But helmets no doubt saved some.

The NTSB studied motorcycles and pedestrians before cyclists and found all three categories of the dead had one thing in common:

“Vulnerable road users are those who share the road with motor vehicle drivers, but do not have the occupant protection that is required in such vehicles,” Sumwalt said.

“They are the road users who are most likely to lose life and limb on our roads and highways. And of all modes of transportation, highway transportation is the mode in which the lion’s share of deaths and injuries occur.”


When it comes to cyclists, some of the deaths are clearly attributable to their bad driving. The NTSB figures would indicate about 30 percent of the approximately 2,400 deaths from 2014 to 2016 involved cyclists failing to yield to traffic or losing control.

But the most common fatal accident involved a motorist overtaking a bicyclist at midblock and killing him or her. The NTSB did not report how many of the drivers involved in those accidents were on their phones or texting.

Neither did the NTSB report provide any hint as to the size of the texting-while-driving threat to cyclists although other data on texting and driving indicates a growing problem.

‘Each day in the United States, approximately 9 people are killed and more than 1,000 injured in crashes that are reported to involve a distracted driver,” the CDC reported in 2017.

“Texting while driving is especially dangerous…,” the agency noted. “When you send or read a text message, you take your eyes off the road for about 5 seconds, long enough to cover the length a football field while driving at 55 mph.”

The NTSB report did note that a study of 5,266 bicycle/motor-vehicle crashes in four states in 2017 revealed that serious or fatal injuries were “twice as likely in midblock crashes” as in other crashes and “65 percent more likely in areas with 30-35 mph posted speed limit compared to 25 mph or less.”

When a motorist takes his or her eyes off the road, the faster the vehicle is traveling the more distance it covers, and the greater the difference in speed between the motor vehicle and a vulnerable road user the greater, in general, the injuries to the latter.

The NTSB’s solution to this problem as regards cycling was a call for separated bike lanes and improved intersections, which some cities have begun building.

“We have identified ways that bicycle safety can improve, and the organizations that can make the improvements happen,” Sumwalt said. “The question is whether these organizations will act.”

The NTSB staff report said 35 states “reported recommending (separated bike lanes) but only four states had them installed along their state roadways.”

The NTSB also suggested making bicycles more visible with lights and reflectors, although the research there is even more conflicting than that on helmets;  using adaptive headlight systems to allow drivers to better see the road without blinding oncoming motor vehicles, which might also save some moose in the state of Alaska; eliminating large-vehicle blind spots, which is being done with cameras in some parts of the world; putting collision avoidance systems in all new cars; and developing “connected vehicle technologies.”

Such technology, according to the Department of Transportation, “would enable cars, buses, trucks, trains, roads and other infrastructure, and our smartphones and other devices to “talk” to one another. Cars on the highway, for example, would use short-range radio signals to communicate with each other so every vehicle on the road would be aware of where other nearby vehicles are.

“Connected vehicles could dramatically reduce the number of fatalities and serious injuries caused by accidents on our roads and highways. While the number of people surviving crashes has increased significantly thanks to airbags, anti-lock brakes, and other technology, the USDOT is shifting its focus from helping people survive crashes to preventing crashes from happening in the first place.”

If cyclists could be wired into the system to make their lives safer, it would be a first for DOT because that claim of rising numbers of people surviving crashes thanks to motor vehicle technology only applies to people in cars and trucks.

For vulnerable road users, technology appears only to have made their lives more deadly, dangerous and scarier. A 2012 survey of Seattle female cyclists found that 79 to 84 percent of them had seen enough distracted drivers on the roads to cause them concern about cycling.

Others are just left frustrated and maybe a little angry.




















9 replies »

  1. Speaking of distractions…that whole story could have been summed up in two sentences above:
    “There is no argument that bicycle helmets will in some crashes save lives…But helmets can and do save lives, and they could save more.”
    This has been proven true in countless studies across the globe.
    “A major study of bike helmet use around the world from more than 64,000 cyclists has found helmets reduce the risks of a serious head injury by nearly 70%…
    The study also found neck injuries are not associated with helmet use and cyclists who wear helmets reduce their chance of a fatal head injury by 65%.”

    • summed up in two sentences if you ignore the fact that Denmark and the Netherlands, where helmet use is low, are the safest countries for cyclists in the world:

      mandatory helmets for cyclists is about the same as mandatory parachutes for private pilots. the safety gear would save some lives. but they don’t do anything to change the physics of the crashes. big impacts kill no matter what you’re wearing.

      meanwhile, if we really want to save a significant number of lives, not just a few hundred, the data is clear: we require mandatory helmets for motorists.

    • Steve,

      How many of those 64,000 in the study you are referencing were putting themselves at greater risk wearing a helmet that was not adequate for their activity level, or simply took greater risks that caused other injuries that they wouldn’t have suffered?

      When are you going to get your site up and running so you can give us a couple sentences of incorrect information from and links galore?

      Speaking of links, you might want to check this one out…

  2. NTSB has a new ‘shot across the bow’ report out, dinging a self-driving vehicle system used by Uber. The system failed to properly “classify” a jaywalking Elaine Herzberg, and just ran her down. It detected her 6 seconds before impact, but wasted the avoidance time considering alternatives … ‘arguing with itself’.

    This has been a big problem since day one, with self-driving and assisted-driving vehicles. Otoh, it is a computer-challenge simply to reliably *recognize* a pedestrian or cyclist, as such. And on the other hand, both can change their location & direction of travel very abruptly … much more quickly & suddenly than can a vehicle, and well outside the tolerances of these systems.

    To have vehicles on the road connected together and communicating, coordinating how they use the road (and in principle also accommodate cyclists & pedestrians)(two different things), means they must be more or less, pretty-much “self-driving”. To effect such coordination, the cars have to be under the control of computer software.

    The real impetus, direction & goal of self-driving technology is perhaps more clearly visible, in Europe. It’s the economy, people.

    Linking-up vehicles is of interest to Governments, in order to move more traffic on existing roads. It’s of interest to Corporations, mainly so trucking can be done cheaper and faster.

    Make no mistake. Self-driving systems are not being created with the average driver’s point of view in mind. Much less walkers or riders.

    You think it’s bad for cyclists in a world of human-driven cars … the victims of computer-driven vehicle accidents (which yes might be reduced *overall*) appear to be vastly over-represented by pedestrians & cyclists.

    • Craig & all,

      This is the link Craig meant to put in his comment-reply, above:

      The Tesla bombshell almost nobody is talking about, from April 2019.

      This article is about a presentation Elon Musk made earlier this year, claiming that during 2020 Tesla will place one million fully autonomous robotaxis in American cities.

      The problem of course is that Tesla still does not have an operational autonomous-vehicle capability. They have an “Autopilot”, and they have a “Full Self-Driving Capability”. Both of these are egregious deceptive labeling, and Tesla is quite upfront about it. [From their webpage, Google ‘tesla autopilot’.]:

      Autopilot and Full Self-Driving Capability are intended for use with a fully attentive driver, who has their hands on the wheel and is prepared to take over at any moment. While these features are designed to become more capable over time, the currently enabled features do not make the vehicle autonomous. [emph. added]

      Worse, the instrumentation and sensors built into Tesla cars is fundamentally limited at the hardware level. [This of course is not a special shortfall of Tesla, but is bedrock engineering “reality”, applicable to all makers.]

      What are the limitations of Autopilot?

      Many factors can impact the performance of Autopilot, causing the system to be unable to function as intended. These include, but are not limited to: poor visibility (due to heavy rain, snow, fog, etc.), bright light (due to oncoming headlights, direct sunlight, etc.), mud, ice, snow, interference or obstruction by objects mounted onto the vehicle (such as a bike rack), obstruction caused by applying excessive paint or adhesive products (such as wraps, stickers, rubber coating, etc.) onto the vehicle; narrow, high curvature or winding roads, a damaged or misaligned bumper, interference from other equipment that generates ultrasonic waves, extremely hot or cold temperatures.

      Under good conditions all vehicle-control technologies (Tesla specializes in (optical) computer vision, while most others lean mainly on LIDAR) work satisfactorily, as intended. The problem is, good conditions are frequently absent. Alaska, for example, is defined by low-angle sunlight, year-round (if you’re lucky!). Low-angle light creates & accentuates shadows, and blurs the shapes of object alongside the road … like brush, tree boughs, mail boxes, and pedestrians or cyclists.

      Elon Musk is just being Elon Musk. Autonomous vehicles, we think, are somewhere over the horizon; probably not a long ways out of view … and Yes Virginia, Tesla will have to obtain Regulatory Approval. Yes, Local Jurisdictions will have lots of say, likely more than the Constitution guarantees them, which is plenty.

      Clue #1. Freeways. Lots, lots easier on freeways. On any fixed, defined, pre-vetted & Approved route.

      Clue #2. It’s not about taxis, nor about You. It’s about Commerce, Trucks, and Business.

      • That’s scary stuff. People, being people, are not going to read the fine print and, much like all safety equipment, are going to rationalize that the equipment with it’s cool marketing name and description means they can casually go beyond their own safe limits.

        Inventors often think mechanically and don’t consider human reality. I see it in firearms tech all the time, even people who should know better insist that there are mechanical solutions to human behavioral problems. We need to stop trying to “outwit” (or force) human beings into “being safe / responsible” and start emphasizing early and repeated training in safety, so it becomes second nature.

        Safety is a behavioral issue, not a technology issue.

      • Good habits/training is right up there with a good dog, Matthew. Training/conditioning will look out for us, keep us out of trouble.

        It ‘works’ in mass-society, of course (for the 1%), to have folks abdicating their responsibilities … and powers.

        The left worked for many decades to shut down firearms & hunting Training programs. With a new outlook in the country, this might be reversed.

    • Let me address the point Craig made about “data”:

      Uber doesn’t begin to have enough data to power the AI. thus it can’t program the computer to make the sort of microsecond decisions at which computers excel. Tesla, however, might have the data:

      Tesla, because they have a large fleet of cars placed with owners, is able to collect camera images (and other information) about all the routes that their customers drive. That’s a big asset for sure; it helps their software, their AI cogitate on a range of issues, and it simply gives them a leg-up on the competition, whenever one of their cars drives into an area they already have recorded.

      But this data mainly helps with the “route”, and the driving of it, and less so with ephemeral features encountered along the route, like people walking or bicycling. It helps ‘some’ to simply gather more images & data of hard-to-spot & difficult-to-predict ‘real world’ pedestrians and cyclists, but these subjects remain tough challenges.

      You may have heard that software intended to recognize people, for example to screen for terrorists at airports, might use the person’s gait to perform identification, rather than try to ID them on facial features, etc. Part of the reason for that approach, is simply that even under relatively well-controlled (indoor & ‘staged’) situations, it is fundamentally ‘shaky business’ for even high-powered AI with massive data and massive computer-hardware on-site, to ‘call’ the identification of people without both false-positives & false-negatives.

      If you place a cyclist in the middle of the road, it is at least then easy to tell that there’s something in the road. But when the cyclist is on the shoulder, the vehicle-cameras are looking sideways off the road, and trying to pick the rider out of a shifting patchwork of ‘moving’ natural vegetation, terrain and the shifting shadows of these features … you’re looking at a real & persistent problem.

      False-positives are often the ‘conservative’ way to lean. When we’re looking for dangerous people, security-wise it’s ok if we stop & question a few innocence people … nicely. But if a vehicle trying to drive itself down the road thinks a mailbox might be a person, and then the rippling of its shadow due to relative-motion during close-approach indicates the ‘person’ is stepping into the vehicle’s line-of-travel, then the AI may well be obliged to either drive you into the ditch or into oncoming traffic (since you are protected inside the vehicle).

      This is a tough problem, a difficult call, and versions of this issue arise in many, many contexts. This hasn’t really been affirmatively brought to the general public’s attention, understandably. However, both the companies who would deploy autonomous vehicles (and be liable for mistakes), and Local Jurisdictions who have to run for election by the locals, are going to be very concerned.

Leave a Reply