One of the great promises of driverless cars is a huge reduction in the number of deaths and serious injuries on our roads. This is because these vehicles will be able to stick to the road rules and react much faster than people can, avoiding the human error that is generally the cause of crashes and their resulting injuries or fatalities:

Superhuman driving skills

Once a computer model is created, then it’s loaded into the car’s brain and hooked up to the rest of the car’s sensors to create real-world model of the car’s environment.

The car uses this model to make decisions about how it should respond in different situations. And because the car has sensors all around it, it has access to a lot more data than a human driver to help it make those decisions.

“There are insightful factors that get factored in. For example, if you are driving along and there’s a parked car with nobody in it, the vehicle will proceed next to that car,” Shapiro said.

“But if it sees the door is slightly open and there is somebody in it, well the expectation is that the door will open at any moment and someone will to try and get out of that car. So at that point, when the car senses that, it’s either going to slow down, or switch lanes if it can, and proceed with caution. And because it has a full 360 degrees view around the car, it can be tracking multiple objects, with much greater things happening, with much greater accuracy than any human.”

Given the huge number of people who are killed or injured from vehicle collisions each year, safety could benefit enormously from driverless technology.

While the occasional mistake will be made by these vehicles, the argument goes that good programming of the driverless cars will slash the number of injuries and fatalities. This article discusses the opportunity in detail.

But I say “could”, because there are going to be some very interesting trade-offs between safety and other benefits that we hope to see from driverless cars. These are covered off in this excellent recent article:

I think it’s important to understand that this outcome for our self-driving car future isn’t self-evident or guaranteed—it is going to take deliberate effort and vigilance because there are already forces at work guiding our streets toward a different, less utopian outcome.

As a matter of practical realism, self-driving cars can and will be safe only directly in inverse proportion to their marketability. That is, you can either have a safe autonomous vehicle, or you can have a mass-marketable one that drives the way contemporary human American passengers and drivers want them to drive, but not both. The reason for this is simple: humans are unsafe drivers, and the customs of the roadway reflect that.

This is something that I have mulled over for a while. If we demand an extremely high level of safety from our driverless cars (as we should), then how attractive are they going to be in dense urban environments. If people know the vehicles will stop safely, which is great for reducing the carnage on our roads, what’s going to stop pedestrians from stepping out in front of these cars all the time? This will be great for pedestrians and cyclists but possibly less great for that other promised outcome of driverless technology, using driverless cars to increase road throughput. In fact, it may well be that sticking to the rules makes being in a driverless vehicle pretty unpleasant.

Exhibit 1: the self-driving Uber that killed a pedestrian in Arizona earlier this year had its emergency braking programming disabled to “reduce the potential for erratic vehicle behavior.” Here’s an analysis of what might have happened if the safety feature were enabled.

Self-driving cars that don’t drive the way a human would drive have a “potential for erratic behavior.” They’re prone to, say, driving under the speed limit, or slowing, stopping, and avoiding potential hazards that a human might disregard. That’s erratic in relation to the current culture of the road, which is to out-drive drivers’ realistic reaction time, use the vehicle’s mass and speed as a threat to keep the way clear, and assume things will all work out.

Modern car- and road-culture expectations assume unsafe driving. When people are forced to contemplate what safe driving really looks like, the way a properly programmed robot would do it, it’s not going to sell cars. Nobody will want to own one—not if it’s going to stop at every intersection for pedestrians, even though that’s already the law. Not if it’s going to give bicycles sufficient following distance and no less than three feet of passing buffer, but also not unsafely cross into the lane of oncoming traffic, instead waiting patiently for a safe opportunity to pass, even though that’s all already the law. Not if it routinely drives under the speed limit to avoid reasonably foreseeable hazards or to adapt to road conditions, even though that’s already the law.

The inevitable, and somewhat scary, implication of this is an upcoming battle between manufacturers and regulators over safety. There will potentially be a lot of incentive for manufacturers to tilt the regulations away from safety:

The profit incentive to disregard laws and regulations is going to be just as strong, if not stronger, than a human driver’s incentive to engage in the same unsafe and socially malignant behaviors to get where they want to go quickly. Only, its going to be baked, universally, into the design of the vehicles themselves. Unless maybe autonomous car owners will get to choose whether to “be safe” or “get me there quickly” to match their own personal ethical calculus.

When a manufacturer decides it’s too “erratic” to stop for pedestrians who want to cross at an unmarked intersection, or too unacceptable to drive at speeds slow enough to be safe—and the car that doesn’t do those things sells better—will existing law prevail, or will the allure of the market bootstrap unsafe driving into a new law of the road, hard coded by car manufacturers to best suit their customers?

This is why we can’t take for granted that autonomous vehicles will be safe, but we will have to insist on it, repeatedly, and in many forums: legislative, regulatory, product engineering & design, and public opinion. As has already been amply demonstrated, compliance with existing law and local preference is not a given with all the financial incentives at play in this developing industry.

While this is a new problem in some respects, it also has many similar aspects to the last time we had a massive transport technological disruption – with the widespread uptake of cars in the early 20th century. At that time there was a major negotiation over street space between cars and others (especially pedestrians). The priority we give to cars these days is so baked in to our sub-conscious that we sometimes don’t realise how it came about in the first place. Peter Norton’s Fighting Traffic gives an excellent overview of what happened. Kent wrote about this book back in 2012:

In a nutshell Fighting Traffic tells the story of how traditional streets where changed from a  place of social exchange, slow movement, and even children playing to become dominated by car movement. The fascinating part of the story was that this transformation did not happen without a concerted effort by the automobile industry. When first introduced to American cities, cars and their drivers were considered pariahs and held accountable for their actions. Local newspapers articles, cartoons and editorials depicted the automobile as a purveyor of death which included macabre imagery such as gravestones, weeping mothers, and the grim reaper. This was easy to understand since even as early as 1930 as many as 16,000 people were killed in automobile accidents every year, mostly pedestrians.

During the early days, automobiles were considered incompatible to city life and this concerned automakers. Through their local automobile associations the car industry went about to change the responsibility of road safety by placing the blame on pedestrians. This was achieved by public media safety programs, school indoctrination, and even street theatre. The most effective method for shifting the public perception was by ridiculing pedestrians. One way this was achieved was through introduction of the term ‘jaywalker’ which implied derisively to a country bumpkin-type (a ‘jay’). Any person not fully respecting the new ruler of the road would be given a card by local Boy Scouts describing their action as inappropriate, and branding them as a jaywalker. This action had the effect of changing social norms, so that eventually laws could be introduced to codify this new value system.

We’re already starting to some similar moves with driverless vehicles, walking back the promises of safety, such as suggesting we should just accept deaths and injuries are a part of the process.

Just like in the 1920s and 1930s, over the next few decades we are heading for a major renegotiation of our streets and will need to confront some pretty interesting questions.

  • Will we accept that a priority on safety and “sticking to the rules” will make travelling in a driverless vehicle pretty unpleasant?
  • Will we place greater priority on speed and comfort of travel, and accept frequent killings?
  • Will we try to get around this problem by physically separating driverless cars from people, even if that ruins our streets? (think fences and barriers everywhere)
  • Will all of this make the introduction of driverless cars just too hard?

I don’t think we can know the answer to any of these questions now, but there are huge risks that we might not end up with safer and more pleasant streets, which means we will need to be vigilant in ensuring driverless vehicles are well regulated.

Share this

74 comments

  1. One other area auto cars could be slow is doing a U turn on a country road. In the weekend GPS said go 4.6 km to turn around . I went 200m and turned around.

  2. Fantastic thought-provoking article Matt, thanks. Continuing the thought experiment for the person in a safety disabled car who sees a kid race onto the road they may struggle to intervene. Even if they see what is happening. There is evidence of auto pilots who in a crisis are so out of practice they cannot do the correct manoeuvre quick enough to get out of trouble. (Tim Harford- Messy). Following this analogy, driving modern aircraft is pretty dull. So would be monitoring an automated car. You are right that safety should be regulated strictly. Passengers can switch off and do something else while bikes and buses can go flying past at rush hour. Plus ca change…

  3. I am still firmly of the belief that driverless cars are a solution looking for a problem, and the sole impetus for their development comes from billionaire libertarian technology evangelists who have a deep ideologically dislike of “19th century” public transport.

    1. Sole impetus? Nah. There’s way too many companies and money involved for AVs to be driven by a ideological agenda. McKinsey estimate that approximately 25 billion p.a. is spent on acquisition of firms with goods/services in areas related to AVs. That’s too much shareholder capital being spent to be explained by an ideological whim. Of course it may still all come to nought, although it won’t be for a lack of money / effort. My sense is that the conversion to AVs will happen, albeit 1) slower initially than many boosters think and 2) faster in the medium/long run than many sceptics appreciate.

      1. IMHO, the “switch over” point (if it happens) will occur rapidly if/when insurance companies start giving a better rate to driverless cars and a much higher rate to traditional cars, and authorities start banning cars from autonomous only driving zones like urban motorways. Human controlled vehicles will then ironically become yet another marker of extreme wealth.

        I see a place for semi-autonomous vehicles in restricted areas with a lot of external inputs available. For example, a car that you can drive to town, drop you off then park itself in a driverless only car parking building that needs far less room for cars to park (since in these new buildings no doors will need to be opened, or passenger access required, and a central computer will control traffic movement). Or perhaps an ability for the car to take over on a crowded motorway where a motorway control computer can interact with traffic.

        1. I would say the rapid switch over will be driverless taxis/ubers. If a driverless car costs $100k and the only other costs are cleaning and electricity, the fares could be pretty cheap and still make a good profit. To make a 20% ROI you would only need $50 a day in fares + electricity and cleaning costs – sounds like a great investment!

      2. I see AV’s solving the ‘last mile’ problem. You’ll be able to get your car to drop you off at a transit station and then make its way home to park without you. It’ll come to pick you up from the station in the evening. No need for park and rides, just a pick up drop off area. Some people will keep their AV’s to themselves and some people will no doubt try to make some money by using their AV as a supplementary transport service.

    2. P.s. The problem AVs help to solve is that of chaffuering. Note that a significant proportion (say 5-10% of travel in urban areas is associated with people driving other people around. Young people and other dependents being the obvious example. I think AVs will initially soak this market up, while at the same time making taxi servives significantly cheaper, say by a factor of 2-3, which will stimulate demand. In the long run I’d expect AV market share to sit somewhere between 20 — 60% travel.

      1. I’d expect that +20 years, AVs will be 50% market. I expect that in +50 years it will be 99% of the market.

        1. I was including travel by all other modes, hence why my upper bound is lower. I think there will remain a residual rump of travel undertaken by existing cars and then all the other modes like walking cycling etc.

      2. a lot of those being driven around are children, especially on the school and daycare run, as kids under 7 years must be restrained, I don’t see how driverless vehicles are going to help here, or if parents will willingly surrender their children to a robot

        1. If these things never crash the kids won’t need to be restrained. Or the parents could belt them in and let the car drive to kindy where someone could unbelt them.

          1. “never crash” is an extraordinarily optimistic statement, so how’s the staffing ratio at a primary school going to fare with even 50 kids in driverless cars arriving at the school gate simultaneously? there’s SUV chaos enough at schools already

    3. I don’t think that’s fair. Humans are notoriously bad and unsafe drivers. I have no doubt that switching entirely from human drivers to robot drivers will save countless lives, all other things being equal. But this article does raise good points that we need to be very careful that safety is kept as a top priority in creating self-driving car regulations. Would should demand zero deaths from our transport systems.

      1. We can’t make safety be the primary goal because our “safety” orientation now leads to perverse outcomes where it’s just safer and easier to ban pedestrians from intersections than giving appropriate facilities. If we only say, “safety!” we’ll get that fenced and herded scenario, similar, but way worse than what we have right now.
        We must insist on not just safety, but also less car dependence and better walking, cycling, PT, i.e., no change from what we should be insisting without AV.

    4. Couldn’t agree more. Driverless cars are the poindextors car fantasy equivalent of the latest overpriced underwhelming iPhone!

      I cannot see these in any safe sense ever because no matter how clever we think they may be nothing is as unpredictable as human behaviour and no road can be as contained and controlled as an airport or rail line.

      Anyway ask Tesla! It couldn’t possibly be their autopilot.

    5. Problem found: we have to pay commercial drivers actual remuneration – screw that.

      I don’t actually think the majority of AV owners will care that their car drives less dickishly than they’re used to. A bus or train ride makes it obvious that passive passengers get far less emotionally involved in traffic than those in direct control of their vehicles. Netflix and chill. Currently we’re in an early adopter phase and passengers aren’t representative of the general public. Similar to when Segways first came out – enthusiasts wanked on about doing tricks on them, when really they’re just a more dignified version of a mobility scooter.

      So bring on the rigid rules and I doubt the opposition will be that fierce.

  4. I foresee reality playing out thus:

    AVs will only be allowed on a model by model basis, with strict safety requirements. The manufacturers probably will ask lawmakers to reduce the safety requirements, due to the performance impacts they impart.

    Lawmakers will initially hold steadfast (fearing their next election and public perception), only to “soften” their view when sufficient numbers of AVs are on the road that they can no longer ignore that they are really slow, due to their safety routines and that the public is favouring convenience.

    Accidents will then happen, resulting in public outcry. Outcry results in safety being beefed up, possibly with legal liability explicitly resting with the manufacturer.

    Accidents will still happen – Ya can nae break the laws of physics, Jim! – The Arizona incident may still have occurred, even if the braking wasn’t degraded.

    AVs will be segregated, most likely by having autonomous mode only allowable on certain corridors or with very strict control that makes it non-viable outside of those corridors.

    Autonomous Vehicles will be a thing, but you’ll still find mostly hands on driving in city streets due to legal liability: Manufacturers enabling autonomous mode in green zones only and lawmakers making it illegal to be hands off outside of green zones or in situations requiring “reasonable care”.

    The real secret ingredient that will make AVs practical for widespread deployment is inter-car communication. The sharing of data “securely” will allow cars to ignore what would otherwise be considered dangerous moves by other vehicles, as they’ve been warned what the other vehicle was going to do by the same. Still only hands off in green zones, though.

    I said “securely” because no security is perfect and eventually somebody will find a way to crack the system, be it for mass surveillance or causing mayhem, it will happen. It’s impossible to determine how society will meet that threat.

    Also, the lead image – If memory serves, that was from Intel wasn’t it? Also if memory serves correctly, they were suggesting that the data load included storage, which is incorrect.

    1. Maybe.
      I see it differently.

      AT may start with driverless trains in their next big fleet upgrade – as per Sydney.

      Kiwirail could then consider it for some of their most predictable routes.

      Auckland International Airport could consider it for some of their shuttle vehicles – again so long as they were highly predictable.

      Then the slightly bigger step will be one of the freight companies such as NZPost adopting driverless autonomous flying vehicles over predictable routes – eg to and from sea port to air port or to Wiri.

      The bigger step will be when Uber or equivalent start to operate autonomous passenger craft – for example from major hotel rooftops to the airport, and back.

      Once the public trust that there are a broader set of areas, operations, and networks in which they already see autonomous transport systems working, they will begin to relax a little about them on the roads beside them.

      Meanwhile, I’m still waiting for my jetpack.

  5. “which means we will need to be vigilant in ensuring driverless vehicles are well regulated.”

    Matt you could do worse than seek a meeting with Minister Genter and request evidence that NZTA have prepared regulations for driverless vehicles on safety grounds – that is entirely within her Ministerial remit to do so.

    FFS don’t let MoT anywhere near it. Nor AT.

    From what I am observing the issue has been bubbling under for a while but the questions of liability are poorly researched let alone poorly regulated. There is no focus within NZTA about it, and they are the key regulators.

    It also needs the motor and life insurance industry – particuarly ACC – to show that they have woken up.

  6. I think there would still be significant safety benefits even if AVs were able to move as cars do today simply through faster reaction times.

    My guess is we will end up somewhere in the middle. AVs will make big safety strides by always obeying traffic lights, pedestrian crossings and by being careful around schools etc, while at the same time we will accept some risk to people who step out without looking.

    I also wouldn’t be surprised if there are significant gains in terms of accurately identifying pedestrian behavior through Image analysis in the next 10 years.

    1. Ever been called Pollyanna, Jezza? Why wouldn’t Jevon’s paradox apply to safety? Increasing safety and all of a sudden there’ll be pressure to:
      -increase the speed limit (It’s not like we don’t see a push for higher speed limits here on GA ‘for roads that are designed for it’)
      -programme the car to go closer to things before reacting
      -design intersections for modern vehicle reaction times, ie less safe.

      I think we know enough about human nature by now to be realistic about this.

      1. No never been called that and have no idea what it means!

        I definitely see your point and it is certainly a possible outcome. However, as we have advanced as a society we have generally become less tolerant to death. We have advanced medical care, vastly improved workplace safety laws etc.

        Even in the area we have lagged behind – road safety, it is still far safer to be on our roads than it was in the 1960s, 70s and 80s. Despite people’s resistance there are many places in NZ that have a lower speed limit than they did 20 years ago.

        In addition to this we are very intolerant of death caused by systems, machines or big companies, so I anticipate we will be much less tolerant of AVs than we are of individuals that are currently responsible for driving cars.

      2. Not to mention hacking the software to “hot rod” a vehicle. But I digress. My guess is that rather than implementing fixed software programs, car makers will allow choosable options. Get me there the fastest way, get me there the safest way, get me there the most economical way, for starters. What about passing a slow moving truck? Get me past this obstacle? How many occupants would use the get me there the fastest button instead, and forget to reset it when past the truck? What about the need to escape a pursuer? Get me there ignoring the speed limit?

    2. “while at the same time we will accept some risk to people who step out without looking” – but surely the car would have to stop if someone stepped out wouldn’t it? Are you really saying the programming could allow the car to run someone over even if it is possible to stop?

      1. No, I’m more thinking of the situation where someone steps out in front of an AV that doesn’t have time to stop. It would still do a better job of braking than a human would but the laws a physics can’t be avoided.

        A truly safe AV would have seen a person walking towards the road and slowed down, a less cautious AV might have assumed they were going towards their parked car and not slowed down, only braking when it realised they were crossing the road.

        1. Good point. But assuming there is room for the car to stop, the car will have to stop (as I guess a human driver would do now if they aren’t looking at their phone or screaming kids). Whether it will then do what a human driver does – honk, yell, punch, etc – is another question.

      2. Yes, degrees of safety programming will become an issue. We will have to ensure these are set at a good level of safety without being ridiculously cautious. I could see shared spaces would be very hard for an autonomous vehicles to drive through & probably better to have these areas totally carless instead.

        1. I disagree. If shared spaces can work satisfactorily with our current stock of vehicles, (although it appears not currently with all our current stock of drivers!) it would be a serious retrograde step to remove this option for street use simply because autonomous vehicles could not cope. European towns often have truly shared road multi
          -use spaces. Cars wait patiently for a pause in the street football game to proceed, and conversly the players immediatly cheerfully wave the cars through as soon as there is a break in play. It works, residents have access to their houses, deliveries can be done, neighbours have a common social space and children can play right outside.

          1. Do you think a car could handle that kind of “thinking”? Waiting for a wave to proceed? I suspect they could drive through them but at a very very slow dawdle to push through in effect.

          2. It appears that a fundamental problem with autonomous vehicles is their inability to receive and process normal readily understandable human communication such as a cheerful wave to indicate come on through. We should not have to change our normal acceptable behavior to simply make a machine more functional. I would find it unacceptable to have a car just barging through a social childrens game, even at low speed, just because it could not interpret normal human social interaction. Instead of getting friendly waves such behaviour would more unlikely provoke a collision between an unguided rock and a guided car. We as people we are constantly non verbally sorting out between ourselves as to who should proceed first. That is a fundamental of our social society and should be jealously retained.

          3. Grant, your ‘slow dawdle’ comment is possibly the only way AVs will be able to negotiate these kind of environments. What should probably happen is that the vehicle detects the obstruction and stops the passenger then has to authorise proceeding at slow speed (5 km/h max) which the vehicle will do until it has clear road ahead or is forced to come to a stop again.

          4. interpreting human gestures is an area of AI that is getting a huge amount of attention. See kinect games for examples. The Mercedes F015 had concepts for human-AV communication back in 2016. I have no doubt that AVs will be able to safely and efficiently navigate shared spaces.

          5. The question is why you’d bother putting funding into it when we already have people who can navigate through shared spaces. Reducing the number of those in cars, and increasing the number of those out of cars will have an effect on safety that is orders of magnitude bigger than any possible safety increase from AVs.

  7. Every time a driver goes on about cyclists breaking the law I wonder how many times I have been overtaken illegally on my bike? Probably 90% of the time is my guess:

    Before you pass:
    make sure you will be able to see at least 100 metres of clear road ahead of you once you have finished passing – if not, don’t pass
    look well ahead to make sure there are no vehicles coming towards you
    look behind to make sure there are no vehicles passing you
    signal right for at least three seconds before moving out to pass.

    Before pulling in front of a vehicle you have passed:
    make sure you can see the vehicle in your rear view mirror
    signal left for at least three seconds.

    If driverless cars apply those rules they will be stuck behind bikes most of the time.

  8. Autonomous cars must be kept clear of two trends, now detectable in New Zealand, towards greater emphasis on safe walking and cycling.

    The first reason is geometry. Autonomous cars will be able to keep closer together, but not around non-autonomous cars and not very much closer; think long trains and emergency stops. Walking and cycling can do much better than this, at much lower cost.

    The second reason is a combination of fuel scarcity and climate change, and a third is liveable cities. Again, walking and cycling is the way to go.

    The rules need to be set before the cars arrive in any numbers, and doing that internationally will be near-impractical. But programming choices will be practical, so perhaps international agreement is less necessary.

    But if everything is nice and safe, what to do about children deliberately stepping out in front of autonomous cars?

    1. Agreed, any talk of significantly reduced following distances is either boosterism or deluded; ultimately any following distance less than the vehicle stopping distance plus a safety margin relies on far too many single points of failure to be trusted by any serious safety analysis. It is worth remembering that at 100 km/h a car travels around half its length every tenth of a second. There is every chance that AVs will be required to have a ‘black box’ in the future so any dirty laundry will be aired when one of them crashes.

  9. AV’s have huge potential for the bus companies. They could change 30% of their fleet to AV’s and use them for all the NIS buses that drive around our city, or Hunter/Killers as they are better known. They would still need drivers in the other 70% of buses that actually carry passengers.

  10. Thanks Matt, some great points. Google themselves said in esrly trials their cars programmed for full safety mode were less efficient for freeway merges than human drivers who tended to “push in”. A cautious computer would not “push in” at freeway speeds.

    Connected and Autonomous Vehicles (CAVs) promise to reduce this problem, but only for car vs car interactions. Car vs ped is still a conflict of time versus safety. Safety should win from a public interest point of view, but then I think that about stopping climate change too. We still haven’t.

  11. Autonomous road vehicles should only be considered in New Zealand after autonomous light rail vehicles operating on shared roadway space have established somewhere in the world, a proven enhanced safety performance above that of manned LRV’s on the same route. If this cannot be achieved, then there is no chance of vehicles that must manage their own steering function as well, being able to achieve an enhanced safety performance over currently available vehicles. I lay this down as a realistic challenge, and a credible path forward to those currently developing AV’s and those advocating for their approval on our roads.

    1. Yes a key thing is going to be street & urban design that works for the best outcomes. We don’t’ just want the status quo with the vehicles we have now swapped for autonomous ones.

      1. I’m more pessimistic than some, and see these cars reclaiming the streets from peds and bikes because the cars simply can’t detect pedestrian behaviour at all. Expect to see shared spaces disappear, segregated footpaths, and pedestrians forced by law to cross only at light-controlled crossings. We’re giving these machines way too much credit for making appropriate decisions. What Microsoft will never tell you is how many times an update is actually a revert to the previous version because it messed things up. Will it be any different for these cars with lives and not just your silly Facehook access at stake?

        1. I’d probably look more closely at the computer software that already flies aircraft than your desktop PC’s software in terms of reliability.

          The drivers of cars today regularly can’t predict pedestrian behaviour either.

          1. Well, the software about to be introduced for these cars is about equivalent to what the aircraft industry had back in the 1960s. The platform (operating system) is much more mature of course, but the unforeseen circumstances and unintended consequences caused by changes are at about the same level of development.

          2. When was the last time a plane had to deal with a pedestrian crossing the road?

            I don’t think you can compare auto-pilot with self driving cars, the environments are completely different.

          3. Dan – my comment was around software reliability rather than complexity. I agree driving a car is more complex than flying a plane in terms of computation.

          4. Why should we expect plane auto-pilot reliability for something that is far more complex? Greater complexity means greater room for human error. Software developer error that is – not driver error.

            Uber have already demonstrated they cannot reliably detect pedestrians.

            “the self-driving system software classified the pedestrians an unknown object, as a vehicle, and then as a bicycle with varying expectations of future travel path.”

            Teslas have frequent updates to fix software faults, just recently consumer reports in USA found the Model 3 had terrible stopping distances, and Telsa released a software patch to fix it. Why did Telsa not pick this up in their own testing?

            So i’d have to say in summary I am pessimistic given the track record so far.

    2. The automatic soap dispenser in the mens room can’t even tell when i am WAVING MY FUCKING HAND IN FRONT OF IT FOR TEN FUCKING MINUTES>

      …tell me more about autonomous cars

      1. 🙂 OK, here’s the story. There’s a small child in each of that longs to be able to give commands and have the world obey. Lights that come on when we clap. Soap that squirts tidily when we put our hands out. It’s a delicious fantasy, and a little of it makes us laugh, for a while. When we grow up, we realise that there’s more satisfaction to be had by walking that mile with our own feet, planting the tree with our own hands, making our own piccalilli sauce.

        But since people can make money from indulging the fantasy, the big corporates find fancy ways to indulge these fantasies. Taken to extreme, it’s called an autonomous car.

  12. Of course things will be much safer. We just need to get past the tech hurdles. The first cars, planes and trains were crap. The current CAV’s are still under development and are still far from driverless.

    A CAV doesn’t get drunk or distracted so that eliminates a huge chunk of deaths.
    A CAV doesn’t get impatient or break the speed limit or tail gate so that eliminates plenty more of speed related deaths.
    A CAV has sensing capabilities far better than the average human driver and reacts in a fraction of the time, so that eliminates some amount of those “step out into traffic” deaths.

    There will still be death and injury, but it will be much less and remain mostly the fault of poor decisions by humans (currently 80-90% of crashes).

    So what if the car drivers more cautiously. You will be watching Netflix at time. You’ll be wanting the drive to last a few minutes longer and go around the block again. The driverless car turns potentially unproductive time in traffic into potentially useful time.

    The roads will be much safer. And more more congestion than now.

  13. Safety needs to be the top priority, but not the only priority. We can’t go around sacrificing safety for the convenience of pedestrians, cyclists, or PT. users. We got into this mess in the first place by sacrificing safety for the convenience of motorists. Once we can say that no-one is going to die, then we can discuss how to best balance the convenience of different modes. But we mustn’t ever trade lives for convenience.

    Saying that banning pedestrians from an intersection is the only way to make it safe is just wrong. Lowering the speed limit and/or putting in a pedestrian crossing are low cost solutions that sacrifice some motorist convenience for pedestrian safety.

  14. Unless the technology can be retrofitted to existing vehicles at a reasonable cost I can’t see the average person being able to afford them and without widespread use I see them causing more problems than they cure.

    Imagine the average impatient driver behind a self dive vehicle that slows to 50kph at every intersection because higher speeds kill in side impact crashes,they are going to want to overtake at the first opportunity.

    I think Ari is right though, ultimately they could make roads far safer and more congested. I would add, but slower. Think of the boost that would be to building a high speed rail system, that’s if there’s enough electric power left in the grid to operated all this autonomous activity. Jervons Paradox comes to mind.

    Personally I think it’s pie in the sky I don’t see more complexity as the future but less, we will be forced by debt, climate change and the need to cut CO2 to our coats accordingly.

  15. Why must driverless vehicles be cars? Assuming a shared ownership model most vehicles could be small (room for 1 or 2 people only) with a few larger ones for more passengers or carrying bulky items. And combined with driverless trains/light rail for mass transit.

    An autonomous tuk tuk or similar would make a lot of sense for many urban trips, in terms of fuel/energy efficiency, space efficiency, cost of materials produce and perhaps most importantly safety, at least for humans outside the vehicle.

    1. How will these autonomous vehicles clean the urine stains and sneeze residues, and clear out the rubbish, each night? Will the streets be filled with hundreds of thousands of zero occupancy vehicles driving themselves to mass cleaning centres filled with thousands of employees?

    2. My prediction:

      The first big wave will be the rich early private adopters.

      After that will be driverless taxis. Remove the driver and you can run a car most of the day at $0.80 a km + %10 profit. So maybe $1 a km, maybe a bit extra for data. It will drive all taxi drivers out of a job (but not out of business), except for those passengers that prefer a more personal experience. The taxi driver will end up owning multiple cars and sending them out to make money and cleaning them during the day at quiet times and what not. They will be very smart about where they locate their cars for best utilization.

      As public acceptance grows, next you will get the shared mini buses that shuttle along certain busy routes stealing some bus patronage, picking up people who book. Splitting the costs between passengers will easily give them rates comparable or better than the bus, without needing a 50% subsidy from the government. This is a huge advantage.

      I can see old people taking out mortgages on their overpriced houses to buy an AV and make some extra money during the day. Get some actual use out of equity tied up in the house.

      All these private owners might pool their resources, stop competing and share the profits and provide a better overall service to people.

      Big disruption coming in terms of business models.

      The roads will be safer, but congestion will probably get worse.

      Trains will always be needed because they can carry lots of people at peak times and the roads in the CBD can only carry so many cars, driverless or not.

  16. Driverless cars will only ever be a gimmick restricted to a few very small locales where the roads are engineered for them. The most important question of driverless car technology has never been asked.

    Do people want them?

    It seems to me the only people who want to see this technoogy develop are those who are developing it, and those who don’t drive cars.

    I suspect the reason that question is never posed, is simply that the answer is not liked.

    I am quite certain driverless cars are the monorails and jetson cars of our time. Definitely a happening thing.

    1. Geoff.

      If you asked someone if they wanted an iPod or iPhone before they existed, what do you think they would say? Henry Ford supposedly said “If I had asked people what they wanted, they would have said faster horses.”

      You are looking at the problem from a very narrow perspective that this technology is a gimmick. Is guaranteed safer roads a gimmick?
      Most people don’t like their job and would quit if they didn’t need the money. The key here is need versus want.

      If you get down to it most people don’t want a car, but they need one. Most people don’t like buying petrol, but they need to in order to use their car to get to the job to get the money to pay for the petrol.

      Most people don’t like being stuck in a car in traffic 2hrs of the day. Wouldn’t you love to catch up on netflix instead? Voila, here is a AV for you. Don’t want to own a car with all those hassles, but still need to get to that pesky job in privacy and comfort door to door? Voila, here is a shared AV scheme for you. Want a “safe” way to get your son to ballet practice 45min drive away? Voila, here is a AV for you!

      Car companies are not selling cars. They are selling access. Some people like to drive. Most people don’t. The car enables access. People don’t want the car, they want the access the car can give them.

      The big question is whether PT can compete for that same need/demand for access.
      I think it is unlikely, except at peak times. I suspect, if done well, most people would pay a premium to use an

        1. Can a country the size of NZ run both PT and private cars or are we robbing other sectors to pay for our car culture. how much does it cost to duplicate.

          Could we have better schools and hospitals if we didn’t cater for single occupancy travel.

  17. Imagine how easy carjackings will be with driverless cars.

    Also, general “trolling” especially in the inner city, where groups of pedestrians can exploit known behavioural patterns of an autonomous vehicle to cause it to flounder aimlessly. Fear of death or injury can be a good motivator!

  18. Couldn’t the same argument be made for any tech? It would be much faster to blitz up the onion if my food processor didn’t have to have a lid on it. Or the microwave would be more convenient if it didn’t have a door. Access to the pool would be more convenient if it didn’t have a fence around it. Of course there will be regulations, and as we’ve seen with the recent news stories, accidents are extremely bad for business in the AV world. So safety will be prioritised by regulators and manufacturers.

    At the same time, you can’t drive past every pedestrian at 10kph just in case they suddenly jump into the road. AVs run a probabilistic behaviour model which accepts that there is some small probability of error. That error will be (or already is?) much smaller than for human drivers, and in case something unexpected happens, the AV has much shorter reaction times.

  19. “But if it sees the door is slightly open and there is somebody in it, well the expectation is that the door will open at any moment and someone will to try and get out of that car. So at that point, when the car senses that, it’s either going to slow down, or switch lanes if it can, and proceed with caution.”

    This sounds exactly like what manycyclists have to sense, judge, process and act-on already.

  20. Cyclists definitely need more protection from dooring, among other ways they are endangered. What if an AV could look behind it and if it sees a cyclist coming, it blocks the door from opening? (With some kind of beep or warning so the passenger doesn’t think they’re being locked in.)

Leave a Reply