Will Driverless Cars Save Road Cycling?
By Chris Carmichael
Founder and Head Coach of CTS
There will come a day, in the not-so-distant future, when the cars that pass you as you ride down the road will be driving themselves. With the known dangers posed by woefully imperfect humans behind the wheel, it’s tempting to view autonomous vehicles as the savior of road cycling everywhere. The question is, should cyclists throw our support whole-heartedly behind the development of driverless cars?
Will autonomous vehicles make the roads safer for cycling?
In the short term, no. Driverless car technology has come a long way in less than a decade, but it is still pretty primitive compared to what it will likely look like in another 10 years. At the moment, pretty much everyone – automakers and researchers – knows driverless cars are not very good at recognizing or reacting to cyclists (or pedestrians, for that matter). It’s the unpredictable human component they can’t control for.
Cars – even with drivers – tend to stay in their lanes, or at least between the curbs. Experienced, training-oriented cyclists tend to ride predictably, too, but there are plenty of people on bikes – including kids – who are less predictable. It would be naïve, however, to think the technological hurdle of recognizing and reacting to cyclists won’t be overcome within the foreseeable future.
The Benefits of Driverless Cars for Cyclists
Distracted driving is getting worse:
We may someday reach a tipping point where society takes a stand against distracted driving in the same way it did against drunk driving, but we’re not there yet. In the meantime, more drivers than ever are looking at screens instead of through the windshield.
The technology inside cars isn’t helping, either. Centralized touch screens are bad news for anyone outside the car. Tactile control surfaces like knobs and buttons can be found and operated by touch alone, without taking your eyes off the road. A centralized touch screen with multi-level menus inevitably takes a driver’s attention away from the road. And yes, some newer cars respond to voice commands, but I still can’t tell if that’s actually a step forward or just a work-around for unworkable touch screens.
Driverless Cars Require Safer Infrastructure
When I was contemplating this column, I talked to Robin Thurston, whose career has frequently put him at the intersection of tech and endurance sports. He co-founded MapMyFitness, sold it to and worked as Chief Digital Officer for Under Armor, and was the CEO of the consumer genomics company Helix. “Driverless cars will benefit from more physical objects like concrete separators, medians, and bike/pedestrian lanes to better delineate their paths,” Robin believes.
Cyclists have been fighting for decades for infrastructure improvements that could enhance cycling safety. But like it or not, the cycling community has a smaller voice and shallower pockets compared to major automakers, oil companies, and trucking companies. The infrastructure improvements that may help make driverless cars and trucks a reality – including more physical separation between driving and cycling lanes – could potentially deliver the safer cycling spaces we’ve been seeking. Of course, that’s only if the cycling community has a seat at the table, otherwise it’s just as likely that the new infrastructure would eliminate space for cyclists altogether.
Sensors see what drivers don’t
While the technology in autonomous vehicles is still not very good at recognizing cyclists and pedestrians, there are some things we know sensors and cameras can do better than humans. An array of sensors or cameras can look both ways – and in all directions – simultaneously. This solves the problem of a driver scanning back and forth, or just looking in one direction, before pulling out into an intersection.
Autonomous cars use a combination of cameras, LIDAR (Light Detection and Ranging), and radar to see the environment around the vehicle and determine the location and relative speed of objects nearby. Different sensors have their strengths and weaknesses, but together they may prove better than humans at “seeing” cyclists in the dark, through the glare of sunrise and sunset, and in the rain.
Robin Thurston also pointed out that the autonomous car can respond to information faster than a person, and that response is not affected by stress or fear. People sometimes freeze under pressure, but computers don’t.
Driverless cars aren’t angry
An autonomous car doesn’t have bad days or short tempers. The operating system isn’t in a hurry and doesn’t get frustrated by the mere idea that cyclists are on the road. Driverless cars won’t throw stuff at you (at least not yet) or purposely buzz you with a “punishment pass”. In short, an autonomous car might be the most patient and considerate driver on the road, which is sure to drive the irritable, impatient, and late-for-work passenger absolutely bonkers.
Driverless cars respond to the conditions
Just like a driverless car doesn’t get angry or rushed, it also doesn’t take unnecessary risks. When the weather gets bad, like the roads are wet or there’s snow or ice, the autonomous car will respond as it’s programmed to respond to the conditions. It will slow down, or brake sooner, for instance. Drivers, on the other hand, routinely fail to adjust their driving behaviors to changing conditions.
Driverless cars won’t misjudge a cyclist’s speed
In addition to the “I didn’t see him” excuse for hitting or pulling out in front of a cyclist, misjudging how fast a cyclist is going is another leading reason drivers turn or pull out in front of cyclists. Thurston pointed out that this technology is already available – for cyclists. “The Garmin Vario, for instance, uses sensors to tell us how far away a car is, and the technology is just going to get better and better,” he said.
Maybe, driverless cars won’t run away
I don’t know this for a fact, but perhaps driverless cars can be programmed so they cannot flee from the scene of an accident. Just recently in Boulder, Colorado, Andrew Bernstein – a former editor of Bicycling Magazine and current member of the True Communications staff – was left for dead by the side of the road after a hit-and-run from a soulless coward driving what police believe was a Dodge van. Andrew was hit from behind and survived, but was left with life-threatening injuries. Number one, if a driverless car relies on cameras and sensors to “see” the environment, collisions would probably be recorded. But number two, maybe driverless cars could have a feature that automatically reports a collision and immediately pulls over and stops after one.
The Downsides of Driverless Cars
Although the transition to autonomous vehicles offers a lot of potential benefits for cyclists, they have their downsides as well.
In the short term – incremental technologies are making drivers less attentive.
On the path from traditional driver-operated cars to fully autonomous cars, automakers are adding incremental technologies to new vehicles. There are lane assist features that nudge you back into your lane if you start to drift. Your car can tell you if there’s a vehicle in your blind spot, hit the brakes if you’re about to hit something, or adjust your cruise control speed to maintain a safe distance from the car ahead of you. In other words, the make it acceptable to pay less attention to the environment around you.
In the long run, I think gradually introducing these incremental technologies is a strategy for easing drivers into accepting autonomous vehicles. The swift change from a fully driver-operated vehicle straight to a fully autonomous vehicle would be too drastic for people to accept. In the short term, however, it’s just making drivers less attentive, which is bad news for cyclists.
They are still huge, mostly-empty boxes that cause congestion
Driverless cars may not help ease congestion at all, particularly if the vast majority continue to take up 45-60 square of space (the average footprint of cars and light trucks in 2017) to transport one person. In fact, they could make congestion even worse, because the number of vehicles on the road will no longer be restricted by the number of people available to drive them. Trucking companies, taxi and rideshare companies, could increase the number of vehicles in circulation because they’re not limited by drivers’ needs for sleep or to earn a livable wage. Considering that time parked is time a service vehicle makes no money, the incentive will be to keep the driverless car or truck on the road 24/7.
Driverless cars lack human empathy
While I think it’s safe to say empathy seems to be in short supply these days, the fact remains that humans respond to our environment and make decisions based on our own values and experiences in the world. People will typically go extraordinary lengths to avoid injuring or killing another human. People stop to shoo turtles and ducklings off the road. And as a cyclist, I’ve seen riders choose to crash in order to avoid hitting and hurting animals or people.
Autonomous cars will face these moments, and will inevitably have to make a choice between hitting a person or another car. A fascinating article in the New York Times reveals the difficulty of agreeing on those ethical decisions based on cultural differences around the world.
Scientists at the MIT Media Lab created an online game called Moral Machine, which was played by more than two million people in 200 countries over a two-year timespan. The game presented players with scenarios, asking them to choose whether the driverless car would hit object A in its path or swerve and hit object b instead.
The results showed how different cultures reflect different values: “Players in Eastern-cluster countries were more likely than those in the Western and Southern countries to kill a young person and spare an old person (represented, in the game, by a stooped figure holding a cane). Players in Southern countries were more likely to kill a fat person (a figure with a large stomach) and spare an athletic person (a figure that appeared mid-jog, wearing shorts and a sweatband). Players in countries with high economic inequality (for example, in Venezuela and Colombia) were more likely to spare a business executive (a figure walking briskly, holding a briefcase) than a homeless person (a hunched figure with a hat, a beard, and patches on his clothes). In countries where the rule of law is particularly strong—like Japan or Germany—people were more likely to kill jaywalkers than lawful pedestrians.”
The unfortunate reality is that, while it’s not going to be the first choice, there will be a scenario that results in a driverless car taking aim at a cyclist to avoid what someone has decided and programmed as an even less desirable outcome.
Should Cyclists Welcome or Protest Driverless Cars?
On balance I believe that in the long run driverless cars will make cyclists safer and make people more comfortable and confident about riding on the road. When you combine advancements in e-bike technology and the acceptance of e-bikes as a viable form of transportation and recreation, the opportunities for cyclists grow even more.
“ I think the most powerful combinations are when one industry integrates with another, so when all these groups start working together I think it will be transformative,” Thurston told me. “Electric bicycles are going to change the conversation, as more and more people want to ride them. City planners and state road budgets will be slower behind these changes, but it will impact them and their decisions.”
It’s going to take years, and perhaps decades, to get from where we are now to the point where autonomous car technology and infrastructure enhancements are advanced enough to achieve these safety improvements. I think it’s in our best interests, however, to encourage and lobby for the development of driverless cars in order to shorten the time between now and a hopefully safer future on the roads.
Pingback: The Scariest Things Cyclists Fear, And How to Overcome Them - Chris Carmichael
For those who want to know what the state of the art of autonomous driving is, please watch Tesla’s ‘Autonomy Day’ presentation on YouTube. The presentation goes deeply into the hardware, software and the building of the knowledge base that trains the ‘neural network’ to recognize the ‘edge case’ situations that are unusual, yet represent accident risks. I believe that one day, driverless cars will make cyclists, pedestrians, motorcyclists and drivers of small cars much, much safer than they are today. Statistically, autonomous cars can be relied upon to ‘behave’ in much more consistent and predictable ways than humans.
Good article but one aspect that we should consider as well, is that driverless car technology is absolutely dependent upon 5G. Reason being, 5G will allow speeds and bandwidth to permit auto-to-auto communications (or other objects to autos). It is this concept that will enable the most significant safety improvements. As an example, one car slows for stopped or slow-moving traffic and this is picked up by the next and the next. While it offers some safety for cyclists, it also means that as cyclists we may need to have some 5G technology operating while we ride, if we truly want to benefit from the improved driverless tech, especially when the unexpected happens (e.g. parked car opens a door into the bike lane and the bike needs to swerve into the lane to avoid a collision).
In my view, there’s till no better safety approach than increasing the number of bikes and motorcycles on the road, and continuing to improve driver awareness. If nothing else, it will help get the right programming and algorithms built into the driverless technology.
While I agree with your point about vehicle to vehicle communication making the network safer, Tesla is focusing its efforts on the ability of the vehicle to use visual data to base its decisions on. At some point, using vehicle to vehicle comm’s, I suspect we could move toward a zero incident environment.
A very interesting, informative, and thoughtful article!
Like e-Bikes, these autonomous vehicles are coming whether people like it or not. I do believe organizations — like PeopleForBikes, L.A.B., etc. — need to lead the way for us cyclists’ safety!
Here’s two of my concerns with these vehicles:
1. Technology can fail or be hacked (recently reported on the news).
2. Infrastructure, including some cities’ “Vision Zero” programs, truly don’t make the roads safer for cyclists! It just creates different conflict concerns between cyclists and other roadway users — especially right hooks, left crosses, and drive-outs.
I take my own safety into my hands by controlling a traffic lane, when safe, legal, and cooperative to make myself relevant to others! It works, especially when riding faster — including those riding e-Bikes not used to the higher speeds and with less skills. I do also wear hi-viz and reflective wear, and use active and passive lighting. A comprehensive approach!
Thanks again Chris for starting this conversation!
It sure is fascinating isn’t it, Derek?