From Slate:

Autonomous cars have a potentially fatal flaw: They struggle to detect and react to cyclists on the road. According to a January 2017 report by IEEE Spectrum, bicycles are generally considered “the most difficult detection problem that autonomous vehicle systems face.”

… we’re increasingly learning that A.I. can amplify our own biases and human failings. If humans aren’t doing a good job of detecting and preventing vehicle-bike collisions, how can we create machines that do the job even better?

One solution presented …(is) bicycle-to-vehicle communications. Instead of just autonomous vehicles (or all motorized vehicles) on the road being able to wirelessly communicate their position and intentions with one another, bikes would be able to join the party. The proposed technology would be brand agnostic, something any cyclist could affix to herself or her bike. …
There’s one problem: This is cheating. Autonomous cars, out there beta testing on U.S. roads today, can accurately detect other vehicles, pedestrians, even big game charging suddenly across a street. Forcing cyclists alone to strap a sensor onto their backs feels like a crutch, a cop-out. …
The problem is that it requires everyone to take part, which poses several noteworthy financial and logistical questions, such as who pays for this system, how it’s deployed, how it’s enforced, and whether pedestrian and traffic laws would need to be changed in order to facilitate cooperative behavior. (For example, stricter jaywalking laws to ensure pedestrians only cross in places self-driving cars expect them to.)
In this scenario, autonomous car success hinges on a large number of difficult-to-control variables. But if the cars themselves are able to successfully sense and react to their surroundings, from a cyclist taking the lane to a toddler dashing into the street, the only variable that needs controlling is the technology itself. …
Bicycle-to-vehicle communication is a good idea and could be useful in certain scenarios, such as when visibility is low—at night or in the rain—or on tricky, twisty back roads with blind corners. But if cars are going to drive the roads without human help, they need to be able to handle all of the challenges that come with it, regardless of whether they’re wirelessly connected to the world around them.

 

Comments

  1. Heaven forfend cyclists might actually have to pay some sort of cost, whether it is a sensor, or licensing or mobility pricing!

    1. According to transportation economists (e.g. Todd Litman) car users in cities are heavily subsided by users of other modes, especially pedestrians and cyclists. The document linked below (p. 6-15) estimates that when a car driver switches to cycling this saves society $12.99 per day. Since cycling and walking are inherently cost saving, charging extra fees to cyclists and walkers would be unjust and counter-productive.
      http://www.vtpi.org/tca/tca06.pdf

    2. I pay lots to ride my bike. Insurance, maintenance, license fee, gas tax, property tax, cycling clothes, etc.
      To be perfectly honest, my bicycle costs more to ride than my car does.

      1. Everybody pays property taxes directly or indirectly. Those pay for municipal roads whereas fuel taxes contribute toward regional roads and highways. They’re unlikely to cover major infrastructure projects which is why they sometimes get tolled. People who never drive pay for roads they barely use, relative to motorists who put huge demand on them.
        You pay insurance because an MV has the potential to do great harm. (It is not meant to fund roads although certain political parties have been known to raid them.) It is irrelevant to this discussion.
        What you pay for maintenance of your vehicle does not contribute to road construction or maintenance. It is irrelevant to this discussion.
        Licensing fees are small and probably just cover the cost of administration. It is irrelevant to this discussion.
        Your sarcasm shows a profound lack of understanding.

  2. Having a bike carry some sort of small device that sends a signal “I am a bike and not a deer” is not too much to ask if the biker is on a roadway frequently, as opposed to a separated bike lane.
    Even e-bikes might have AV features eventually too.

    1. And if a person rides a bike on a roadway infrequently, and thus doesn’t have a transponder, it is OK to run them over? That is ridiculous.

        1. So people have to take responsibility that something else out there might hit them?
          I think that we all have to get together and say enough is enough. From reflector vests to helmets to blinking lights and then transponders!
          We don’t need driverless cars to kill us. The driverful ones are doing a good job of that already.

        2. Seeing many bikers at night without lights, or folks downhill at 60km/h without helmets, or wearing dark non-reflective clothing, or weaving in and out of traffic, I have to wonder frequently whether people’s brains are in the “on” position when biking.
          Of course cars (be they AVs or human operated) have to be operated properly and roads and bikelanes properly marked (and we have a way to go here in BC) but bikers too have to do their part (and they often do not) !

        3. The difference though is that when someone drives a car and does something wrong it’s considered to be an individual failure whereas when someone riding a bike does something wrong it’s considered that all others who ride bikes are also guilty.

  3. How do you roll out a system of transponders in the first place – make them mandatory and get millions of bike riders using them before putting autonomous vehicles on the streets, or do you start with the cars and let a few crushed cyclists act as an incentive for other bike riders to get the transponders? What sort of coverage will we target – is it okay to risk the lives of 40% of riders? 10%? Inevitably it will be the poor and disadvantaged who won’t have the transponders (and who are most likely to be riding erratically). I wonder too about the liability for car owners & manufacturers if they know that their cars fail to recognize & respond to vulnerable road users.
    We are often told that autonomous vehicles are going to provide a lot more road safety but if they can’t reduce pedestrian/cyclist deaths (or even increase them), then this really limits their ability to live up to their promise.

    1. Don’t worry about the liability. Insurance will cover it.
      No politician is going to start spending money from the public purse to give cyclists tech gadgets for their own safety. That would be too much of a liability.

  4. I agree that forcing cyclists to carry transponders seems morally dubious. But if there is such a requirement, then it seems that drivers should bear the cost of the transponders, as they’re the source of the danger.

    1. People have a responsibility for their own lives and life styles. If it involves biking (or sky diving for that matter) it is UPON THEM to have devices to decrease their risk.
      Cars have more weight and more buffer zones, incl AVs so the risk is on the biker.
      I’ve seen a tomb stone once which said “But he had the right of way” !

      1. You keep telling us we should all pay for the infrastructure to support suburban lifestyles. How do you square that position with a sudden belief in personal responsibility for the costs associated with lifestyle choices?
        The false equivalency between cycling for transportation and skydiving is ludicrous btw.

  5. As a cyclist I’d be ok with it, if it didn’t require batteries changes/recharging on my part.
    If you could possibly be killed because a battery died it isn’t gonna work.
    Working lights I’ll take responsibility for.
    A dead transponder to save me from autonomous vehicles, not so much.

  6. An early reported problem with AV programming wasn’t that they couldn’t detect people on bikes, it was that they detected them too well. A person doing a track stand, ie stopped at a light and balancing on the pedals, rocking back and forth slightly, was interpreted as a moving vehicle and the AV kept starting and stopping. So we know that AVs can detect bikes. It appears that some programmers might not want them to, and are suggesting that the onus should be put on the person on the bike to carry a transponder. That just sounds lazy.
    Figure it out. It isn’t a problem with cyclists, as the post title suggests. It seems far more like a problem with AVs and their programmers, trying to offload responsibilities.

  7. When AVs do arrive they will be coming from all over the planet. They will have different software and different sensors and anti-collision responses. Those entering BC from other provinces and the USA will quite probably have different systems.
    Collisions will happen. It’s the humans that will need to update their brain software.
    caveat emptor

    1. Motor vehicles already come from all over the planet. We require them to meet federal vehicle safety standards here. Why would unapproved AV systems be permitted under this system?
      We know from testing that the AV sensors can detect bicycles, so that will be an easy standard to include in the regulations. What we will have to test are the software routines that either give the cyclist appropriate space (pass), or cut too close and squeeze through in order to speed up the AV trip (fail).

  8. This issue is just one more reason why AV tech is such a long ways from widespread implementation, and why some of us don’t buy the hype. Square peg, round hole, etc.
    The core problem is not with adapting new tech to car dependent cities, but with the city itself. Nothing is as efficient and affordable to society than walkable, bikeable and transit-oriented communities. No amount of Autotopia Tech Love will change that.

  9. If AV technology can’t decide if a biker is a horse, a goat, a standing object or a pedestrian then I’d say the technology is not yet roadworthy.
    Humans are good at pattern recognition. Machines will take a while longer.
    Honestly, I find it hard to believe if this article is actually true. I bet AVs do know very well what’s a bike and what’s a goat. The far bigger question is: what to do when there is a goat or a bike on a narrow road with oncoming traffic.

    1. I find it tremendously concerning that you think it is a big question about what to do when encountering a bike on a narrow road with oncoming traffic. If the answer isn’t clear, then it is time to stop driving.

  10. Meanwhile, back in the States (New York, Feb 5, 2018)
    Crude oil production in the United States is estimated to have reached a record high in January 2018 and reach almost 12 million BBD by 2019.
    Even though automobiles are transitioning to electric powered systems the sales of aircraft by Airbus and Boeing are exploding, with the Asia market climbing fastest.
    To meet that demand, Boeing estimates carriers will need 16,050 new aircraft valued at $2.5 trillion by 2036.
    Only 10% of people in Asia have flown in an airplane, so the coming growth is massive. By 2036 Asia is projected to have 3.5 billion passengers per year.
    Over $1 trillion is expected to be spent on airport expansions by 2069, with about half of that due to be spent in Asia, Sydney-based CAPA Centre for Aviation estimates. In Beijing, a new $12.9 billion airport due to open in 2019 will turn China’s capital into one of the world’s biggest aviation hubs. Bangkok’s Suvarnabhumi Airport is set for a 117 billion baht ($3.7 billion) renovation through 2021, including a third runway. South Korea’s Incheon International Airport spent 5 trillion won ($4.6 billion) on a second terminal as it aims to become “the world’s leading mega-hub airport.”
    Airbus now has a manufacturing plant in Tianjin, China which has delivered 354 aircraft. In 2020 production will increase to 6 aircraft a month.
    As of now all aircraft run on oil. The only plans for any other form of power is in sci-fi mags.

    1. That’s why even our illustrious Prime Minister said “KM pipeline will be built”.
      AVs and EVs not really related though .. as AV is the topic of this threat.
      Like EVs, AVs are far away from prime time for the average user on the average road, as allegedly bike detecting software in AVs is not functioning properly or as good as what any person can discern from a distance oh, there is a cyclist on the road”.
      We shall see. I put my name of the long Tesla 3 order queue for $1000. Maybe 2020 ? Not so much for the rather limiting range or the insane price for a very basic car, but rather for the AV “autopilot” features that I find fascinating. I do not like driving in cities and if a machine can do it: awesome. We shall see.

      1. The PM didn’t approve Kinder Morgan to magically support flight to 2069. He approved an Alberta project that amounts to less than a hill of beans to the country in the “national interest” to “balance environment with the economy” and to fight climate change by building pipelines with weak economic fundamentals.
        In other words, he’s incoherent, is easily goaded by the bombast, arrogance, bullying and condescension of Alberta toward BC, would never do this in Quebec, and is willing to sacrifice 18 federal Liberal seats in BC.
        Now, back to bikes and AVs.

        1. Do some tax math, please. Several BILLION $s of taxes can’t be wrong as it buys as a lot of schools, hospitals, teachers, nurses, public old folks homes, civil servants salaries etc That why it is in the national interest, and even Mr. Trudeau knows as much http://www.cbc.ca/news/politics/federal-cabinet-trudeau-pipeline-decisions-1.3872828 Asia needs oil, and if we don’t provide it, someone else will. Oil is very fungible.
          Plenty of Native support for pipelines, incl the rejected Northern Gateway pipeline http://business.financialpost.com/commodities/energy/we-are-very-disappointed-loss-of-northern-gateway-devastating-for-many-first-nations-chiefs-say Plenty ! This pipeline may now will be built – with strong native support – to an Alaska port and the same waters will be used by American ships providing American jobs: http://business.financialpost.com/commodities/energy/first-nations-pipeline-has-a-plan-to-get-around-b-c-oil-tanker-ban-an-old-gold-rush-town-in-alaska
          As such, much like the myth of the sudden death of the ICE (internal combustion engine) and imploding oil use, AVs too will take far far longer to mature into the real world. As we can see with Uber it is still not here in BC ..

    2. Airbus wrote: “Meanwhile, back in the States….”
      Meanwhile, back in the thread about autonomous vehicles….
      We understand that the planes you write about all have autopilot systems. But they don’t run in to people on bikes.

    3. 2069 …. there goes the planet. But wait, oil is a finite resource, and flight is dependent on cheap oil. By 2030+ we will still see flight, just not for common folk.

      1. Give your head a shake. Peak oil is not before 2040 and likely much much later given booming demand in Asia, S-America and now Africa with soon over 1B people all clamouring to get what we have: cars, fridges, vacations, cell phones, condos, homes, ..
        And much like the alleged “fossil fuel” came from organic material millions of years ago, even today oil and gas can be produced “naturally”. Call it “green oil” if you want : https://www.fool.com/investing/general/2014/10/12/forget-what-youve-heard-oil-and-gas-are-actually-r.aspx

  11. We bike riders already are forced to wear helmets and purchase lights to make it less easy for drivers to kill us. I would be quite happy to purchase a transponder as well as long as drivers offer to pay their fair share of road costs. Research from UBC shows that for each km cycled, the bike rider contributes $0.15 to society while each km driven costs society $0.56. It would be easier to simply distribute these for free, not just to bike riders but to pedestrians, dogs, cats and yes – even goats.

Leave a Reply

Your email address will not be published. Required fields are marked *