Air Quality Screen Shot 2016-01-14 at 1.20.09 PM

Published on January 17th, 2016 | by Michael Barnard


Autonomous Cars Likely To Increase Congestion

January 17th, 2016 by  

Congestion is a serious issue in developed countries, but autonomous cars are likely to be as much a part of the problem as part of the solution. Consider these US numbers from a study published in late 2015:

The top 30 bottlenecks are each responsible for more than one million hours of lost time annually. Drivers stuck on these roads experience total delays of about 91 million hours every year, the equivalent of 45,500 person-work years. The lost value of time to the economy from congestion just in this handful of locations is upwards of $2.4 billion annually.

Screen Shot 2016-01-14 at 1.20.09 PMMany advocates frequently allude to traffic congestion as one of the key things which autonomous cars will help solve, but it’s worth looking at the actual causes of congestion and how likely patterns of autonomous car use will play out in the real world. As this graphic from the US Department of Transport’s 2012 Urban Congestion Trends report shows, 40% of congestion comes from bottlenecks. While there are other contributors, the biggest single contributor to bottlenecks is too many cars on the road. Almost everything that can be done to the roads will drive only marginal improvements, while increases in car miles traveled will have a direct and non-linear increase in congestion.

Typically, autonomous car advocates point to two elements which will definitely decrease congestion, but ignore the factors which are likely to increase it.

Reducing collisions reduces congestion — One major advantage of autonomous cars is that they won’t get in accidents with anything like the frequency or severity of human-driven cars. This will reduce overall congestion, while also seriously reducing the number of auto body repair and paint shops required. But they will still get in collisions. The evidence on this is fairly clear:

The Insurance Institute for Highway Safety (IIHS) has seen a 7 percent reduction in crashes for vehicles with a basic forward-collision warning system, and a 14 to 15 percent reduction for those with automatic braking.

That’s excellent, but the automatic braking is one of the primary collision avoidance tools for cars, and it achieves only 14% to 15%. Further, traffic collisions contribute only 25% to congestion. Solutions which decrease collisions but increase road miles traveled as well could have a larger impact with the road miles than with the collisions.

Screen Shot 2016-01-14 at 1.37.52 PMIncreased packing of cars on the road reduces congestion — One potential benefit of autonomous vehicles is that they can communicate with each other and traffic management systems and permit greater numbers of cars to get through a particular stretch of road in an hour without volume-induced traffic jams. This is hypothetical in that it hasn’t been proven with real-world traffic and real autonomous cars yet, but it has been well modelled for decades, and there is little reason to doubt that in general this is true. The chart accompanying this shows varying degrees of increased flowthrough of traffic above the level at which congestion occurs due to the vehicles have varying degrees of autonomous and cooperative driving. It finds that, with 100% cooperative adaptive cruise control — Teslas with Autopilot that talk to one another — an increase of 100% in the normal volume could be managed. This would have a very large impact if car miles driven didn’t merely increase to consume the capacity.

And that’s where the rub is with autonomous vehicles. There are differences of opinion about how autonomous vehicles will actually operate as opposed to idealistic models.

There are different scenarios on this category and its impacts, and differences of opinion. In the absence of any actual data, let’s look at the drivers of behaviour:

Price point reduces transit and increases congestion — Autonomous taxi or Uber equivalents are presumed to be cheaper than taxis or Ubers driven by a human. Eliminating the labour cost from taxi equivalents will likely reduce their cost to consumers. This would make them a more viable option for more trips. But what trips exactly? The Uber / Lyft experience indicates that they are competing to a great extent with transit and increasing total car miles driven. If they become cheaper, then transit gets gutted of riders more, making it less and less economically viable, and congestion increases. If more trips are taken in single passenger vehicles, more vehicles will be in those bottleneck areas, and there’s a humdinger of a problem with that:

Congestion is a non-linear function, so as a road approaches its maximum capacity, small changes in traffic volumes can cause proportionately larger changes in congestion delays.

That’s right, just small numbers of additional vehicles will create traffic jams in free-flowing traffic. Autonomous cars somewhat mitigate this effect, but that doesn’t matter if autonomous vehicles means more vehicles on the road at the same time.

Further, ride shares have to drive further than someone in a private car would in order for the passenger to make their trip, and will often have to travel through bottlenecks empty in order to pick up their customers or return to their designated drop off points. Completely empty autonomous cars on the road are not likely to reduce congestion while requiring more driving.

car-van-suv-trunk-organizer-groceries-sports-equipment-tools-storage-camping-new_2116113Consumers stay attached to private cars — Many people like having cars and use them as extended backpacks and purses. They leave fitness clothes, files, items to be repaired or returned, purchases, and the like in them as they go about their daily rounds. They are status signifiers to neighbours. They are places of privacy. They are vehicles whose sunk costs are turned into travel conveniences, with weekend trips to skiing and annual road trips south made economical by having a car for commuting. Early anecdotal indicators from Tesla drivers with Autopilot are showing much more relaxing long drives, and the Tesla model has free Superchargers, so long-distance driving is becoming much cheaper and easier with autonomous cars. While many automotive theorists predict radical changes in human behaviour related to car ownership, there’s little reason to believe that people will become rational just because of a new technology. There will be an impact on private vehicle ownership, but not nearly as big as some advocates suggest. And if this is the case, then there is little reason to expect a reduction in total car miles driven, the largest indicator of congestion.

Reduction in family vehicles increases total miles driven — While in general there is little reason to believe that there will be substantial decreases in ownership, let’s explore that model, in case it comes true. Let’s identify a model family of mother who works downtown, father who does piece work at home, and two children in school. In this model, the family’s single autonomous car delivers the mother to work downtown, then swings back and takes the kids to school, then swings back and takes the father to a mid-day yoga class, then unspools this at the end of the day. This does take a two or maybe even three car family down to a one car family. Except it does it by up to doubling the total miles driven by the family, as the car goes two ways. Once again, congestion is strongly correlated to total miles driven by cars, and in this model an empty car is driving around with no one in it a lot of the time and the total miles driven by the family vehicle go up. That likely increases congestion.

Parking outside of the city core increases total miles driven — Right now people who drive to work park at or very close to work. That’s pretty useful because they have access to their vehicle and it stops driving once they reach their destination. Under one model posited for autonomous cars, the car wanders away to cheap parking outside of the core. The problem with this is that it once again puts more vehicle miles on the car which is strongly correlated to increased overall congestion. Imagine a bunch of empty cars on the roads at rush hour trying to get through city streets to cheap parking.

Passenger comfort reduces intersection throughput — There’s another downside to autonomous cars related to congestion. People mostly will want them to be as comfortable as trains, accelerating and decelerating smoothly. Studies have suggested strongly that this means that car behaviours will slow traffic through intersections, increasing passenger comfort at the expense of overall efficiency. This can be seen by the innumerable stories about Google Cars driving like short-sighted, very polite, very timid people who have all the time in the world. So in addition to intersections having a bunch more cars in them, many of them empty of all passengers, the autonomous vehicles, especially the ones with people in them, will be taking from 4% (best case) to 2000% longer to clear an intersection. While autonomous cars increase highway vehicle capacity, they could decrease intersection vehicle capacity by 18% to 53%.

Longer commutes become viable, increasing total miles driven — At present, commuting by car is highly unproductive time. Commuting by transit is also problematic due to lack of power outlets, WiFi, and ability to have private conversations. There’s a model of the future, however, in which autonomous cars provide a private productivity zone, allowing people to live even further from work. Autonomous cars promoting suburban living and long commutes as a viable alternative increases traffic congestion, not the inverse.

Induced demand increases volume — Induced demand says that, after supply of something is increased, demand increases as well. This is a well understood economic behaviour that can be observed in efficiency programs and traffic. Study after study after study shows that when you increase road carrying capacity, the roads just fill up again.

Duranton and Turner say building more roads results in more driving for a number of reasons: People drive more when there are more roads to drive on, commercial driving and trucking increases with the number of roads, and, to a lesser extent, people migrate to areas with lots of roads. Given that new capacity just increases driving, they find that “a new lane kilometer of roadway diverts little traffic from other roads.”

There is no reason to believe that increasing road carrying capacity due to autonomous cars being able to drive more closely together with fewer volume-induced slowdowns will not be subject to exactly the same effect. Ignoring all other aspects of adoption, if more vehicles can get through a specific road segment in an hour, it will probably fill up to that maximum volume with resulting congestion bottlenecks reoccurring.

It’s also worth, of course, pointing out that all of those additional vehicle miles would produce much more carbon and pollution emissions in the absence of any other changes. Thankfully, electric vehicles are leading the charge in this space, and there is a strong likelihood that significant penetrations of fully autonomous vehicles will match the necessary electrification of those vehicles.

Screen Shot 2016-01-14 at 2.14.27 PMSo, the traffic congestion scenario is not as rosy as many autonomous car proponents suggest. For the purposes of discussion, I’ve put together some projections of different scenarios. In each, I’ve reduced collisions by 40%. Most other factors don’t change, bottlenecks due to volume swings substantially with increased vehicle miles and the non-linear effects discussed earlier.

At best, the greater number of cars which would be able to get through a given point without congestion ensuing would be consumed by the greater number of cars both full and empty driving through that point on average. This is behaviour which is remarkably similar to building new roads, in that relief from congestion is always temporary. “If you build it, they will come” is the reality of traffic. In this scenario, there is still some benefit from reduced collisions, so overall traffic is better.

In the median scenario, all benefits from reduced accidents are consumed by greater numbers of vehicles on the road, leading to exactly the same amount of congestion as before, just balanced to more cars rather than more collisions.

And in the worst case, there’s a potential for much worse congestion due to all of the vehicles on the roads.

As I look across all of the systemic implications of autonomous cars, I see equal or greater congestion on our roads, not less, and more societal negative externalities than advantages. Autonomous vehicles are not a silver bullet for traffic problems. They are coming, and as a person who drives, I welcome them, but good policy and adaptation to the transformation requires clear eyes.

Drive an electric car? Complete one of our short surveys for our next electric car report.
Keep up to date with all the hottest cleantech news by subscribing to our (free) cleantech newsletter, or keep an eye on sector-specific news by getting our (also free) solar energy newsletter, electric vehicle newsletter, or wind energy newsletter.

Tags: , , , , , , , , ,

About the Author

For the past several years Michael has been analyzing and publishing reports and articles on decarbonization technologies, business models and policies. His pieces on electrical generation transformation and electrification of transportation have been published in CleanTechnica, Newsweek, Slate, Forbes, Huffington Post, Quartz, RenewEconomy, RenewablesInternational and Gizmag, as well as included in textbooks. Third-party articles on his analyses and interviews with Mike have been published in dozens of news sites globally and have reached #1 on Reddit Science. Much of his work originates on, where Mike has been a Top Writer annually since 2012. He also has published a climate-fiction novel, Guangzhou Future Tense.

  • neroden

    I look forward to the *safety* benefits of robot cars — they’ll be less likely to drive on the sidewalk and run over pedestrians —

    But we should have no illusions:
    — in the city, they’re going to be slightly slower than human-driven cars
    — they will not reduce congestion in the city *at all*: trains will still be needed in cities to speed you past the traffic
    — they won’t work on the rural roads in the countryside, which are a practically impossible target to program for (they never correspond to the map, have poor striping and signage, no shoulders, dirt or gravel surface, etc): manually driven cars will still dominate the countryside.

  • eveee

    Thanks. Great input. My input? Why don’t cities do the logical thing with bike lanes? Make them protected with barriers or curbs. Logically, pedestrian sidewalk, curb, bike lane, curb and barrier, curb, parking, traffic lane. Its dumb to stick a painted bike lane between the car traffic lane and the parking. I don’t buy that cities don’t have money for it. They put in structures for traffic calming and beautification all the time. Just turn them into a better function. As you say, there are clear benefits to the bike lanes. We just need to be able to make them work properly. It needs real physical barriers, not plastic or paint to do that.

    • neroden

      That’s called a “parking protected bike lane”, bike advocates have been advocating for that design for several years, and I don’t know why it isn’t completely standard.

  • Doug

    This article really misses the point that autonomous cars may drastically reduce the number of “owned” cars as people shift towards ride sharing services. In the example of Uber, they already include congestion pricing. It’s not a huge leap to determine that if households drop from 2-3 cars to 0-1 cars it’s not physically possible to have more cars on the road at the same time.

    • No, that is covered in two different points which indicate that the impact on ownership will be less than some think and the impact of reduced ownership on congestion will not be what people expect but more likely the inverse. There’s a lot of fuzzy thinking about actual impacts and a lot of narrow studies which don’t take a more systemic view in this space.

    • neroden

      Given that most cars sit parked all the time, it is absolutely possible for there to be lower rates of car ownership AND more cars on the road moving.

  • ROBwithaB

    This is interesting, thank you.
    I’d like to see some more detailed studies, but I think the finding is clear and supports the intuitive assumption that, the more widely the system is adopted, the cheaper it becomes. Route planning is complex if you only have a few users in any particular area. But if you’re able to include almost ALL commutes, you suddenly have a LOT of flexibility to mix and match routes.

    • kent

      Exactly, and with a larger pool of passengers, you need proportionately fewer cars idle per square-km to cope with random surges in requests, in order to satisfy those requests within 1 minute. As a result, the simulations show that as the pool of passengers increases, time spent idle and vehicle-distance-travelled-per-passenger-distance-travelled goes down and occupancy per vehicle in peak periods goes up. That is, as the passenger pool gets bigger, the system goes more efficient and costs per passenger km fall.

  • There are too many holes and assumptions with this article that I just don’t have the time to debunk. I’m stunned that this article made it to CleanTechnica.

    • Bob_Wallace

      Your comment wastes your time and ours as well. If you have problems with the article then list them.

  • jnffarrell1

    No need for mass production to clean up NYC.

    One use case is ripe for google-bugs. As few as 3000 google-bugs could fully instrument and control traffic at the 25 mph limit from Ninth Avenue to First Avenue and from 32nd Street to 52nd Street. No taxi scofflaw could threaten a pedestrian, much less another driver, within a half mile radius of Grand Central Station.

    Ticketing could be automatic and one could easily add tickets for noisy mufflers and honking.

    Mixing use cases and production goals muddles thinking about driverless cars:
    1) Premium Cars will never be truly mass produced.
    2) Mobility aids for old people and the handicapped must be low volume and personalized.
    3) Police cars will be low volume and contain police-specific driving-aids….and so on. There is a whole lot of muddled thinking going on.

  • Richard Poore

    What many of the comments seem to be overlooking is the transition between what we have now and whatever version of autonomous driving we have.

    Autonomous cars can drive closer… to other autonomous cars. Human driven cars are the problem in the mix, and will be for a long time. Currently autonomous cars cause accidents because they dont drive like humans. As the auto car control improves (mimics human compatible behavior) we will see fewer accidents but congestion problems from the human/ auto mix are going to still be a concern.

  • Ivor O’Connor

    Nice article in that it brings a lot of aspects into focus. The conclusions though are wrong.

  • ROBwithaB

    Would you exchange your daily commute in your personal vehicle for one in a spacious, comfortable, shared (driverless) vehicle, if such shared vehicle offered a door-to-door service, could be summoned 24 hours a day, and saved you 90% of your monthly transport bill?

    I think I probably would.
    People tend to think of autonomous driving in the context of the vehicles we already have. But the big advantage of the tech is that it opens up the possibility for vehicles that do not yet exist. Something like a very spacious 8 or 10 seater. Entrance via the front or rear. Literally, the whole front of the vehicle could open up for you. (No driver to worry about.) Corridor down the middle. A row of big comfortable chairs on either side of the aisle. Overhead bins. Internally the layout is almost like a small regional airliner. Do this all on a “skateboard” platform. Without engine, driver or drivetrain, you have a surprisingly small footprint, probably similar to existing SUVs on the road, and a very spacious interior layout.
    If these autonomous minibus “roadliners” had access to dedicated bus lanes, and could ride in “trains” with other autonomous public transport vehicles

    The other possibility that autonomous driving opens up is for proper real-time synchronisation of passenger needs with vehicle availability, coupled with instant response to traffic conditions. A “swarm” of interconnected vehicles will respond far more intelligently than any individual vehicle, with or without a driver.

    I might yet keep a vehicle in the garage for emergencies, but most of the time it will be MUCH simpler, cheaper and more convenient to summon something to take me where I want to go.

    • neroden

      Autonomous buses are a much more likely application than autonomous cars in the long run.

      They have the advantage that they always follow a fixed route, which makes the problem of programming to deal with “the unexpected” a lot simpler.

      I hope someone figures out how to use autonomous buses in rural areas (because rural roads are HARD for autonomous vehicle design). It would be valuable, because rural bus service with drivers is totally unaffordable, so the cost of rural living is high and demands a car.

  • I think the effect of increased packing that increases road capacity is played down.

    In real life, human driven cars are separated by a distance of about 1 – 1.5 seconds (well below the advisable safety distance of 2 seconds). Due to the limited reaction time of us mortals. There is nothing to prevent communicating autonomous cars to drive bumper to bumper. At a speed of 100 km/h, this increases road capacity by a factor of 5 or more.

    The idea that intersection throughput is reduced due to increased demand for comfort is of course rubbish, and the % mention are totally pulled out of thin air. Mike ignores that today’s traffic at intersections is slowed down by class 8 trucks and other types of slow traffic. And the ‘normal’ cars? I’m not sure in what city Mike lives, but most human drivers are incredibly slow to get off the mark or take turns so slow that I can hardly imagine an autonomous vehicle to go even slower. On top of that, humans perform all kinds of unexpected manoeuvres, greatly impeding a smooth flow of traffic,

    Autonomous cars would communicate with traffic lights and adapt their speed to arrive at the intersection at the right moment (when the lights turn green), not to needing to stop at all. A platoon of bumper-to-bumper autonomous cars would pass in a matter of seconds. Compare that to the time needed for a bunch of human drivers simply driving up to the red light, stopping, waiting and then having to pass the intersection when the lights turn green.

    But I agree that the vehicle miles will increase very much and undo many of the advantages. But will the net result be negative? I doubt it. But as long as human drivers are still a significant factor, it will probably deteriorate, only to improve after human driven cars are outlawed.

    • Jenny Sommer

      There will be a lot if non autonomous vehicles mixed in for at least another 30years.

      The one thing autonomous cars don’t do is slow down and watch who might got killed in some accident.

      • newnodm

        There may be no autonomous vehicles in the mix for 30 years.

    • mike_dyke

      I totally agree arne.

      Most (all?) traffic jams are fundamentally caused by people having to wait for something to happen – a gap in the traffic, traffic lights etc which causes a knock-on effect to the following cars having to wait until they’ve got out of the way.
      With autonomous vehicles, they anticipate more quickly and alter speed smoother so the jams go from “stop-start” to “flowing slowly” which improves the arrival rate at the junction which caused the problem. If you can get people arriving at a junction with nice gaps, then you can zip/merge the two lanes together easier hence reducing problems at that intersection.

      After that, you look at road layout to improve matters.

    • Bob_Wallace

      Having experienced many occurrences of people entering the intersection after the light has turned red if I’m the lead car I usually wait until I’m sure the other cars are stopping before I start. Clearly I’m adding a bit of lag to the flow of traffic. Autonomous cars in communication with the other cars would have no need to hesitate.

      • Two points. The first is that the reference points to expectations of smoothness for the passenger being the cause of delays. The second is that cars share intersections with bicyclists and pedestrians, and more caution due to poorer communication will be required.

        • Bob_Wallace

          I hesitate when the light turns green because I’m not in contact with cars that could enter the intersection from the right or left of me. I wait until they have declared their intentions by significantly stopping. (I was almost killed by someone t-boning my car.)

          Autonomous cars can communicate. They can also monitor all pedestrians and bike riders at once.

          Once all cars are autonomous they can move through intersections in one block units, operating like a coupled train. All start and accelerate at the same speed.

          • neroden

            Nope. Wrong. I’ve explained this up above. You’re just plain totally wrong.

            Look at the actual implementations of autonomous trains (now working since the 1970s) if you don’t believe me.

            “Platooning” has been the dream of autonomous car lunatics since the 1970s, but it’s fundamentally unsafe and will never, ever, ever be implemented on real roads.

          • Bob_Wallace

            Do we have any systems now operating in which vehicles are in communication with each other? Or are you taking data from ‘dumb’ vehicles and assuming conditions will be the same for ‘smart’ vehicles?

            If vehicle B is following vehicle A and A is broadcasting its direction, speed, intentions, and a “all systems fine” signal then vehicle B can operate differently from a dumb vehicle which has to wait and see what A does before reacting.

      • neroden

        Failsafe design for an autonomous car has to assume that the other autonomous cars are lying or broken when they claim that they won’t speed into the intersection. (What if the brakes are out on that other autonomous car and the car computer doesn’t know it?)

        So you don’t get much of an advantage really; your autonomous car still has to wait to *see* the other car stop. The shorter reaction time is really the only advantage, and that’s not much of an advantage.

    • Pulled out of the air is an interesting way to describe a study referenced in the article. To quote the gloss from CityLab:

      “The final traffic tolls ranged from annoying to frightening. In the baseline situation, without any driverless cars, each vehicle experienced a delay of 20 seconds at the intersection. When driverless cars accelerated and decelerated in the style of light rail, the congestion worsened from 4 percent (21 seconds) to 50 percent (30 seconds). The number of cars traveling through the intersection—at 1,793 in the baseline scenario—also fell between 4 percent (1,724 cars) and 21 percent (1,415 cars).

      The HSR-smoothness scenario was even scarier. Against the same baseline, autonomous cars that started and stopped like high-speed rail increased delay anywhere from 36 percent (27 seconds) to nearly 2,000 percent (6 minutes and 44 seconds!). Meanwhile, intersection capacity fell between 18 percent (1,469 cars) and 53 percent (850 cars).”

      Similarly, autonomous cars aren’t magic and can’t magically eliminate physics. They will be significantly better and safer, but assuming 100% reduction in collisions is unrealistic. 40% reduction is extraordinary and much to be desired. Assuming anything more is intellectual laziness.

      Apologies, but your unreferenced dislike of the material and references provided is not convincing.

    • neroden

      A failsafe design for an autonomous car has to ALWAYS assume that the car ahead of it has broken down and will stop dead without warning. It must leave a safe stopping distance.

      Therefore, no, autonomous cars can’t be driven bumper to bumper. Sorry to spoil your fantasy.

  • Autonomous passenger drones. In effect multi-level freeways. Now we’re talking.

    • MarTams

      Flying drones to pick up and drop you off. Then we don’t need to pay for toll and road maintenance.

    • ROBwithaB

      I see no reason not to incorporate small passenger drones into an automated multi-node multi-mode public transport system.
      But expect to pay a lot for the privilege.
      Because physics.

  • dirtysquirties

    It’d probably help if people would quit living so far away from work. It seems like once or twice a week I hear a bunch of coworkers complain about their 1-2 hour, 50+ mile commutes. Sometimes they sound proud of themselves when they get to boast about how much more miserable their commute is than the other guy who complained.

    I understand a lot of people have legitimate reasons for not living as close as possible, but it baffles me when people who have a choice would rather waste a huge portion of their lives sitting in traffic than have that free time to themselves…

    • Bob_Wallace

      It would help but people are going to have long commutes for a variety of reasons. A nicer house, some green space for their kids, better schools, one partner works “50 miles away” and the other does the commute, they’re stuck in a house and the market is down.

      What’s needed are solutions. Much better public transportation with autonomous driving cars on each end. Or autonomous driving cars so that people can make something out of their lives while commuting, even if it’s catching a nap.

      A few days ago I rode the MegaBus from Sacramento to San Francisco. Comfortable. Free wifi. Charge outlets. Everyone get a seat, no standing. $5. That one worked for me. Cheaper than what it would have cost me to drive and I didn’t have to drive/park.

      Once I got to SF the options were not as smooth. I paid more to go from the Bay Bridge to the airport than for the first leg. Had to change from train to BART. And if I wasn’t going to a common spot I’d have been in a cab. Public transportation needs to hub more efficiently. And autonomous driving cars would make the last mile a piece of cake.

      • MarTams

        The cities of yesteryears were originally designed around mass transit. Places of residence, places of work, places of goods and services. It took few minutes of walking from place of residence to get your groceries. But when car manufactures became big and influential, they got their hands on city planning and were against encouraging of designing cities around mass transit and the use of mass transits. They wanted for each household to have separate home and cars, the great American Dream. And so here we are in this mess, where we waste a lot of energy for commutes.

        • Bob_Wallace

          I don’t think we need to lay it all on car manufacturers. People wanted to live outside dirty, noisy cities and cars gave them the ability.

          Developers, not car companies, designed the suburban neighborhoods. People wanted to move there for the quality of life.

          • neroden

            No. It was partly the car companies directly. It was partly the lead poisoning in the cities (which was the fault of the car companies for putting lead in the gasoline starting in the 1920s).

            And it was partly Big Government, which starting in 1945 subsidized suburban tract housing, and in the 1950s started to prohibit people from building new housing in the cities. I’ve gone through the history in an incredible amount of detail. There was a lot of top-down government interference to prevent people from living the way they wanted to — in the city.

            “Quality of life” stuff was not the main cause. Though the lead poisoning meant that there was a temporarily valid quality of life concern — which is almost totally gone now, apart from places like Exide.

            There were, of course, suburbs built for “quality of life” before the lead poisoning. They’re called “streetcar suburbs” or “railroad suburbs” for a reason, and they’re totally differently designed from modern auto-dependent suburbs; they’re also super-expensive becaue they’re so popular.

      • Freddy D

        “What’s needed are solutions. Much better public transportation with autonomous driving cars on each end”. The route you described is a perfect example – Sacramento to SFO cannot be helped by autonomous cars because demand for that route far exceeds what 1,200 cars per hour per lane can provide. Perhaps by a factor of 20, so adding a lane or two to highway 80 wouldn’t be enough. Even doubling the throughput because autonomous vehicles could run much closer together safely wouldn’t be sufficient. Furthermore, that would double the cars arriving and parking at both ends, and SF is now so saturated that it is reducing cars, not adding them. Finally, even Megabus runs at 70mph on the best day, and on that route in reality will come to full stops from time to time.

        Your solution: old fashioned high speed rail, which can carry up to 20,000 people per hour per lane in each direction, never gets slowed down by congestion, runs at well over 100 mph (a route like that might be regional rather than full 200 mph intercity, I don’t know), and autonomous vehicles become a last mile solution. That can scale. And travel speed matters.

        LA is another example – autonomous vehicles can perhaps optimize the freeway system, but LA needs solutions that fundamentally deal with the fact that cars take space and the city is full of them so there’s no room left, regardless of the skill of the driving software.

        • Bob_Wallace

          There’s a difference between improve and solve. Autonomous cars would improve road conditions, tighter packing and fewer accidents. No slowing to rubberneck.

          With spontaneous car pools traffic would lighten.

          I’m not pushing a cars only solution. Never have. We need to move more people to mass transit, IMO. I’m really hoping the Hyperloop works. If people can travel at speeds higher than flying, in more comfort and for far less money then we’ll see cars come off the roads. Give people cheap rides on each end via autonomous EVs and few people are going to bother driving long distances.

          Parking. With autonomous driving cars where one simply pays for the miles ridden there’s no need to retrieve any specific car from the garage. You’ve seen the car ferry wheel? Try a 30 story version in an industrial area. Pack thousands of cars into a small space.

          • neroden

            Hyperloop won’t really work. Alon Levy has debunked it.

            There are real physical advantages to a pair of conical wheels on a pair of steel rails. It’s passively self-stabilizing (unlike Hyperloop or similar schemes) which means you can make very long trains (unlike Hyperloop or similar schemes, and unlike road vehicles) without worrying about jackknifing (unlike tractor-trailers or articulated buses) at high speeds with minimal infrastructure (just that pair of rails).

            The limits to high speed rail at the moment are mostly acceleration limits set by passenger comfort. Hyperloop doesn’t change that.

          • Bob_Wallace

            Alon Levy wrote his paper in 2013. I’m sure hundreds of engineers involved with Hyperloop projects have read it carefully.

            The second Hyperloop test track just began construction. By the end of the year we should have some proof whether the ‘loop works or doesn’t work as both test tracks should be in operation.

    • Jenny Sommer

      Working and working too much is generally a problem. Reduce the workload to 5-10h a weak and you will be much happier.

    • neroden

      People live far away from work because it’s illegal to build more housing next to work. It’s due to zoning codes. Specifically, something called “Euclidean zoning” (named after a court case involving the city of Euclid).

  • BigWu

    Did they count the Chris Christie administration’s Bridgegate-induced gridlock under ‘Work Zone’ or ‘Special Events’?

  • Peter

    Interesting article. One thing not mentioned is the impact of smaller vehicles. If shared autonomous cars becomes dominant I think it’s likely that the typical car will be a small electric single seater. They might then be able to run two abreast per lane and double road capacity through that alone.

  • Otis11

    Did you think about the changes that will occur with public transit as well?

    Think many more 15 passenger EV-vans that travel the routes more often (every 5 minutes instead of every 15), or travel smaller routes that couldn’t have justified a bus.

    One popular objection to public transit I get is that they have to travel quite a ways to the bus stop, wait too long for the next bus, and it’s rarely on time.

    Combine this with a good app that allows users to communicate needs with the system and you could have fully personalizing public transit. Think about a system like google maps where you put your current location and destination in and it plans a vehicle to take you straight to your destination – door to door service (with maybe a few stop to pick up/drop off others along the route). If the route is too far, it brings you to a central hub where you transfer to a bus/train/subway (and it can coordinate with the bus/optimize drop-off order of other passengers so that you never miss the bus) and has another van waiting at the other end to take you from the bus/train/subway to your final destination.

    With improved public transit, vehicle ownership (and use!) may decrease significantly. While the fears presented in this article are valid, I don’t think we know enough to accurately quantify which effects will be larger. Only time will tell.

    • Jenny Sommer

      Never gonna happen. As complicated as it can get.

      • ROBwithaB

        Not so complicated if you can automate the entire process. Machine learning, swarm computing, real-time price elasticity and route optimisation. All pretty easy with modern computers.
        Including the ones that most of us already have in our pockets.

        • Jenny Sommer

          I wish I could share your optimism. This doesn’t only have to work but people need to use it.
          The problem isn’t the technical implementation but the user/costumer.

          • ROBwithaB

            Agreed. But if a system is simple and convenient AND it saves the customer a lot of money, then SOME people will start using it.
            And if it works really well for those people, they will tell their friends, who will start using it and will tell THEIR friends…..

            One only has to start with small pilot programs. As long as they’re well integrated into existing public transport infrastructure, the convenience will be obvious to anyone using the system.
            It really is just a software problem.

          • Bob_Wallace

            You’re right. IMHO.

            If people can ride where they want to go, in roughly as much comfort as their own car, not have to drive, not have to deal with parking, and do it all for less money than owning/driving one’s own car many people will move over.

            We don’t need everyone to move out of personal vehicles. Just enough to unclog our roads and cities.

          • ROBwithaB

            As I pointed out in another comment, it’s a virtuous circle.
            As soon as you can get SOME people using an “automated, driverless, J-I-T price-elastic last mile minibus swarm” public transport system, the efficiency improves, allowing for more frequent scheduling and lower costs. You can start with a single route, and expand from there. The more routes, the greater the ridership, the greater the system efficiencies.
            It incremental advantage goes round and round until your marginal unit cost per trip drops to pennies. Some (rich) people will still drive cars, but will pay dearly for the privilege.

      • Carl Raymond S

        And if you went back in time just 30 years, and told people that they would soon search all of the world’s news/videos/photos/archives with just a few keystrokes, from in their home or on a beach, and have results in less than a second, they would have said:
        “Never gonna happen. As complicated as it can get.”

        • Jenny Sommer

          But I am still right on flying cars and hoverboards 😉
          And Musk will never sent humans to Mars in his lifetime.

          • Carl Raymond S

            I’m a little more, or less, optimistic (depending on whether you are the chosen astronaut). I don’t believe we will see a human return from Mars in our lifetime.

          • Bob_Wallace

            Mars’ gravity field is 38% of Earth’s. 60% less fuel to get up high enough to get on board a ship that can move people back to someplace like the Space Station. Then take public transport home.

          • Jenny Sommer

            Count me sceptic on the one way suicide mission also 😉

          • Carl Raymond S

            Apparently there are thousands of volunteers. Some would do it just for the twitter followers.

        • Jenny Sommer

          I’ll take back the ‘never’ on customized puplic transport routing…make it a ‘not in the next 20 years’.

        • neroden

          Bullshit, Carl.

          I was around 30 years ago — that’s only 1995 — and I absolutely said that yes, I would be able to search most (not all!) of the world’s news/videos/photos/archives from my home and have very quick results. (Actually, by 1995 I already *could* do most of that: there was just a lack of common interfaces slowing down the search through different archives.)

          It was, in fact, bloody obvious in the 1980s to anyone using ARPAnet / NSFnet or the other predecessors of the Internet that this was going to happen.

          • Carl Raymond S

            OK, so it was clear to you (and I maintain it was not clear to most) in the 80s that the amazing power of Google would emerge, yet you are sceptical now that autonomous flexible route cars/minibus will replace scheduled fixed route buses?

            I’m not seeing why one was likely while the other is not. They are both progressions of computing power, networking and software development. While the date is variable, I feel it’s a question of when, not if. If I were a driver or pilot with more than 10 years to retirement – I’d start re-skilling myself now.

      • Otis11

        Uber and other such drive sharing services are already starting to try things like this… what’s to make you think a city won’t adopt it?

    • ROBwithaB

      Let us think for a moment about buses….
      What is the basic problem with buses?
      Well, they’re too big, for a start, which makes it difficult for them to
      negotiate narrow streets, and they need their own dedicated bus stops
      to avoid blocking traffic, which means slowly pulling out into moving traffic each time they load or offload passengers. They block up intersections. They actually end up causing traffic congestion due to their size.
      They stop all the time, and each stop takes too long.
      They’re slooow, even when they’re moving.
      Unless you have a bus stop right on your doorstep, there’s usually quite a long walk to get to the designated bus-stop. (And if there IS a bus stop on your doorstep, you often learn to hate the bus because its noisy, as are the people getting on and off and waiting for the bus.)
      You usually have to wait too long for one to arrive. If you just miss the bus, the next one can take forever to arrive.
      And then three all arrive at the same time.
      And the first one is always really crowded while the last one is empty.
      (If you have some curiosity and a LOT of time, simply google “why do buses all arrive at the same time”. It’s a real thing.)
      They suddenly stop running at 10pm or whatever, and only start again at 6am. (or, if there is a “night bus” service, they come along once every hour, which generally means you may as well walk…)

      These all sound like different problems, but are in fact all the result of the SAME particular problem.
      Buses are simply TOO BIG.
      For a sixty seat bus, you need perhaps 30 people at any one time to make a route viable. You need a lot of people waiting about before it makes economic sense to serve them.
      Which means that buses run infrequently, and only on the major routes, and have a limited number of high volume stops, and stop running completely when demand is lower than a certain threshold.
      And WHY are buses so big? Well, because you need a driver, whose time is valuable. One of the biggest expenses in the system is the salaries for the drivers. So in most places, where labour is expensive, it is uneconomical to run smaller buses.
      As soon as you get rid of the driver, it suddenly makes sense to run six small buses of ten passengers each, every five minutes, rather than one big bus of

      sixty passengers once every half hour. And you can afford to have more diverse routes, with sub-arterial feeders. And because you’re never inconveniencing more than nine other people, and because any person on the bus is only going to be subjected to about 6 stops (on average, someone with better knowledge of stats can help me out with this) you can basically stop and load/offload anywhere that it’s safe. And each stop will also be much quicker than with a bigger bus, because there’s fewer people getting on and off, and it’s easier to merge back into the traffic.
      It’s a virtuous circle.
      As buses run more frequently, it becomes more convenient to take the bus, so more people use them. As ridership on the system increases, the density of trips in a particular area increases, incorporating more side streets into the system until almost door-to-door service is routinely possible, meaning even greater convenience and increased ridership, allowing for even more efficient route planning, leading to lower costs, leading to increased ridership and even more frequent buses,, getting even more cars off the road and decreasing congestion and speeding up commute times, leading to lower costs, etc, etc.

      At the moment, public transport is terribly inefficient. There are many ways to make it MUCH more efficient. Whoever does this will make a LOT of money.
      Write the software and license it to municipalities all over the world.

      • Otis11

        Exactly my point (but explained much more thoroughly than I)!

        I would note, there are a few places where the big busses make sense (running from a central hub in the city to a more remote hub in the suburbs), but putting smaller vans to pick up and distribute people on either end would make it so much more effective!

        • ROBwithaB

          I just typed a few thousand words, explaining the concept for an entire citywide public transport system, a hybrid of scheduled and on-demand modes, and then Disqus crashed on me. (Perhaps nature’s way of telling me to get back to work).

          To paraphrase:
          Yes, big buses, trains, trams etc will always be useful, even necessary, in an efficient, integrated public transport system. One needs those high volume, high speed bulk movers to do the “heavy lifting” that is required when you have millions of riders a day. The “hub and spoke” system works perfectly well for goods and packages, no reason that it can’t work for moving people about too.
          It makes a lot of sense to accommodate the “frequently traveled” mainline trips with bulk movers. Two random people might well have a commute that is exactly the same over 80% of its length.
          It is just the 10% of the journey on either end that is different.
          That’s where the challenges come in, and that’s where driverless vehicles (or various configurations) can offer some very attractive solutions.

          • neroden

            I think of that “last mile problem” as being mostly a driverless taxi problem. If you’re at the “boondocks” station, that’s going to be the most popular solution. At an “urban” station you walk to your destination.

      • neroden

        No. The basic problem with buses is that they are caught in traffic.

        The solution is trains. Like it has always been.

        People have studied this a lot.

    • neroden

      Autonomous buses aren’t going to be allowed for a long time for legal liability reasons (sigh), but the fact is that autonomous trains already exist and work just fine. And they’re a lot cheaper to operate than trains with drivers.

  • Steven F

    “The Insurance Institute for Highway Safety (IIHS) has seen a 7 percent reduction in crashes for vehicles with a basic forward-collision warning system, and a 14 to 15 percent reduction for those with automatic braking.”

    This data is for human controlled cars and is not directly relevant to the subject. Especially when there is data from the current experimental autonomous cars is available. Most of the accidents recorded involving autonomous cars were not caused by the computer. Most were due to human drivers hitting the autonomous cars. If all cars are autonomous there would be very few accident. the crash reduction could easily exceed 70%.

    In an autonomous future many people will choose not to own a car. Instead they will sign up for an autonomous car service. The autonomous car starts its day picking up a passenger and then driving him to work in the city and hour or more away. The car then picks up another passenger nearby and drive her to work only a few miles away. The car then does a couple more small trips to work. After doing several more short trips the car parks itself and plugs into a supper charger. After getting a full charge it does several more trips before starting to pick people up from work and bringing them home. In this way the one autonomous car eliminates the need 4 or 5 cars or more.

    In my personal experience when there is a minor holiday (only in which only government and bank employees observe) Traffic moves a lot smoother. I can even tell when schools are closed. The number of cars on the road is not a lot less but you don’t need to reduce traffic a lot to speed up commutes.

    “There’s another downside to autonomous cars related to congestion. People mostly will want them to be as comfortable as trains, accelerating and decelerating smoothly.”

    I have notice today that there appears to be more and more drivers that only have two driving modes, full throttle and break. They almost never maintain a steady speed or cost. Those people that drive efficiently accelerate more smoothly then hold a steady speed and if they see a red light a head they will cost and then gradually break. Increasing 0-30mph acceleration time 1 or 2 seconds is going to feel a lot smoother but it is not going to make a big difference in clearing the backup at the light. Also cars could travel faster and closer together on the road with improved safety cancelling out much of the slower traffic light through put.

    The Author is making a lot of assumption on how autonomous cars will be used . Many of which may or may not be true when autonomous cars become available..I strongly suspect that they will have an overall positive impact on traffic.

    • Jenny Sommer

      Why should I not want to own a car when it is cheaper than now in the first place? What are you going to do with kids seats and all the stuff you leave in your car?

      • Bob_Wallace

        You may want to own only one car per household. Not two, three, ….

        • Jenny Sommer

          Didn’t consider that because I never thought about owning more than one car.

          I guess empty vehicles would try to avoid getting into or even causing congestions.
          But if there was a permanent bottleneck that even autonomous cars couldn’t circumvent most of the time this would be a rather practical approach to road planning.
          Here at university they use fungus for that..

  • Bob_Wallace

    Riding mass transit during rush hours sucks. Many people will spend money and effort to avoid being jammed into a box with too many other people.

    Autonomous cars would allow for very convenient car pools which would simply pick one up at their door and drop them off at their other door. No problems with one pool member not being ready on time, needing to change their schedule, whatever. A rider would just input their departure time and destination and the system would optimise how best to pick them off and deliver them.

    Cars would turn into comfortable mass transit. More expensive than higher capacity systems but cheaper than single occupancy. Moving from one person per car to an average of two cuts congestion in half.

    • Freddy D

      Riding well designed mass transit during rush hour may suck from the crowding standpoint, but it does get you there quickly and on time. Stuck in traffic, even in a self-driving car would require massive padding of the schedule because you’d never know when you’d get there.

      Plus, self driving car doesn’t solve the problem of matching up carpool buddies, where there are trust issues, safety issues, compatibility issues on music, listening to radical am radio talk shows, etc.

      Finally, well designed mass transit gets 30,000 people per hour through a one lane corridor, where freeways get 1,200. 2,400 if you double people up.

      • Bob_Wallace

        ” the problem of matching up carpool buddies, where there are trust issues, safety issues, compatibility issues on music, listening to radical am radio talk shows, etc.”

        There are workarounds.

        Safety – passenger camera, “get me out of here now!” cell phone button, “don’t ever put me in a car with the person sitting in the right rear seat again” button. It’s not like one person could take over and drive the car somewhere else. And they would be totally identified by their contact/billing info.

        Music – no vehicle sound system, people can use their own devices.

        • Freddy D

          How the vehicle gets piloted doesn’t fundamentally change any of these issues. These issues are fundamentally the same as today and people haven’t been able to figure it out, even with all the smartphone apps in the world and the gift of free time in the form of HOV lanes. Yes, some folks do manage to figure it out, but the incentives for carpooling (HOV lanes, let alone cost-sharing) have been compelling for years, and it remains a small percentage. There will be some new apps coming out to try and match people up, but I predict continued gridlock of single-passenger cars.

      • Carl Raymond S

        I don’t think it’s an either/or situation. Autonomous EVs will support mass transit. Buses routes wander about – train routes, not so much.

        Sydney has trains which carry 2000 passengers. A lot of bus routes, like mine, are ‘feeders’. I live 3km from the station, and the bus stop is 600m up a steep hill. It’s a pain to walk/bus/train to the city because the intervals are 10mins/25mins/35mins, plus a 10 minute wait for the bus (rarely on schedule) and a 5 minute wait for the train (reliable and regular).

        Yes, you read that right, the bus takes 25 mins to travel 3km because it takes the most indirect route imaginable in order to cover the zone and fill the bus. It loops back on itself twice!

        An on-demand driverless EV acting as feeder would change the dynamic completely. 10+10+25min would reduce to perhaps 15 mins, because it would pick me up at the bottom of the drive and we’d only need to make a few pick ups which were actually on-the-way to the station. That’s an hour less commuting per day. Yes please.

        • ROBwithaB

          There is presumably a “sweet spot” that the market will decide upon. Depending on the type type of land use and the commuting patterns, it would probably make sense to use really small one-person “pods” in certain areas at certain times of the day. In other situations, a 10 seater “minibus” would be much more efficient, if there were a large number of people sharing a very similar commute. In South Africa, we have a great number of privately owned minibus taxis which provide an (almost) door-to-door service. The system isn’t terribly efficient at the moment, and I have often thought of ways to make it more so.

          The widespread adoption of smartphones now makes it possible. If one could let the “swarm” know where you are and where you want to go at what time, it would be possible to allocate resources very efficiently. Especially so with autonomous vehicles. Kinda like the ride-sharing apps, although I have my doubts as to the long-term viability of that business model in its current form.
          I can foresee fleets of autonomous electric vehicles, wholly-owned by municipal governments or locally licensed “taxi companies”, directed by generic software systems. The fleets would comprise a number of different vehicles sizes, from the smallest of “pods” to large buses or even trains/subways/trams etc. With autonomous vehicles, one could effectively create “trains” of minibuses in the HOV lanes of major arterial routes.

          The way it might work is this: I log onto an app on my phone, and using a map interface, or a pre-selected list of my usual travel start/end points, I tap in the desired route.
          The computer responds almost instantly with a number of options. These would include various combinations:
          a) a door to door service in a single vehicle, with a single occupant, at a specific guaranteed time (most expensive)
          b) same as above, but with a few minutes of flexibility on the pick-up time
          c) single vehicle multiple occupants door-to-door, but with a number of stops to collect additional passengers along an optimised route, with estimated journey time posted.
          etc etc all the way down to:
          d) multiple stops entailing changing over to different modes, and a certain amount of “final mile” walking or cycling, and an elasticity of perhaps 30 minutes on initial pick-up time (Very cheap)

          So my cheapest possible option might entail something like this: Tap my route into the app, indicating I don’t mind a bit of a walk and a bit of a wait. After perhaps 15 minutes, I get a message to let me know that the minibus will be coming in ten minutes. I put on my coat and grab my backpack etc and walk 100m to the bottom of my quiet residential street, to the intersection of one of the major streets in the suburb. The minibus shows up within a minute or two and picks me up. I greet the four other passengers. The minibus stops four more times along the major feeder route, for perhaps 20 seconds each time, to (almost) fill the seats to capacity. We all travel a further kilometre or two and are dropped off at the metro station, where the next train arrives in 5 minutes (and already knows that I will be arriving, so I know that there is space available because I’ve been “booked all the way through”. I travel on the metro for 20 minutes and arrive at a downtown station. There are a number of buses waiting, and I am directed to the number 7 bus (which is expecting me). I can find the status of the bus to see whether I have time for a quick bathroom break, or a cup of coffee, etc.
          Ten minute bus trip takes me to my peri urban office node. At the bus stop there is a small fleet of minibuses waiting, which do round trips around the ring road of the office node. I’m dropped off about 100m from the main entrance of my employer’s bland corporate monolith, and walk briskly to make sure I arrive within the half hour flexi-time window that has been allocated to me.
          App might them automatically send a message to the pre-programmed number for the wife to tell her I’ll be leaving the office half an hour late. I would also be prompted to select the same route in reverse for the evening commute, and given a pick up time. And an alternative (later) pick up time too.
          Very similar to the standard commute in a large metropolis, but quicker, a lot more convenient, and likely to be a lot cheaper too.
          You can customise any particular part of the route to suit your own needs, and get a “quote” one the spot for any combination of nodes and modes. Would also be very easy to see how much you could save by doing a part of the trip by bicycle or on foot.

          The biggest problem with public transport at the moment is the (lack of) integration between different modes and nodes. If “the system” knows before-hand the details of your ENTIRE trip, it is much easier to make efficient allocation of resources, especially it is knows the details of every other trip too, and can shift the lowest-priority trips by maybe half an hour or perhaps an hour either way.

          If you were about to buy your “monthly commuter pass” from the muni, and were offered a 70% discount to travel each leg half an hour later than normal, would you take it? Not everybody could, obviously. But enough people would do it to allow for a much more efficient use of the entire system.

          The biggest problem might well be the privacy of commuters. I’d be reluctant to use a system like this unless I could do it anonymously.

          • Carl Raymond S

            Wow Rob, you win the essay of the week award. There might be a venture capitalist reading this who will fund that app… or steal the idea and cut you out.

            A few thoughts I had while reading. Scheduled v unscheduled is a topic unto itself. Buses are scheduled, taxis and our hypothetical on-demand minibus are not. There are times when you just can’t get a taxi – midnight on new years eve for example. Where services switch to unscheduled, people have to accept that in the main they will be faster, but will occasionally be slower. A net win, but an occasional loss. I don’t think this is unworkable – very few people lose their job these days for occasional lateness, especially if they mostly arrive (and start working) early.

            Where services are scheduled, as anything with more than 20 seats will have to be – I don’t see how they can reserve a seat for you, just because you pushed a button on an app saying you were prepared to pay some extra percentage. That has echoes of a class based system that supposedly died out last century.

            I like the idea of virtual trains. Not sure what happens if a duck or child wanders onto the road. Tyres and roads are not uniformly grippy. Even if they brake in unison, the gaps won’t stay constant (or maybe they can all stop at the same rate as the worst). Somebody will work out the safe distance between train ‘cars’. I assume the train splits if caught mid-length on a red signal.

            My step-son refuses to connect his Opal card (Sydney equivalent of London’s Oyster card) to his credit card out of loss of privacy paranoia. It’s true, big(ish) brother does know where we went and when. I put my faith in government incompetence and under-resourcing. They don’t have the funds or actually give-a-damn enough to track us.

            Banks, google, linked-in, supermarkets on the other hand – they do have motive and means – but they already know everything about us… so the government is a minor concern.

          • Bob_Wallace

            “Where services are scheduled, as anything with more than 20 seats will have to be ”

            No. The “system” will be aware of typical travel patterns and pre-made arrangements and put a one, four, twenty, fifty passenger vehicle in place. The system might send out a fifty first and follow it up with a twenty or four as needed.

            It’s all about calling for a cab with the cab shared and the size variable. If you fly into an airport where arrivals are not frequent you’ll see taxis arriving and lining up about the time the wheels hit the ground. No one knows how many taxis will be needed, but based on previous experience some decent approximations can be made. And, if needed, more help can be summoned.

          • neroden

            The taxi dispatching problem is actually absurdly difficult and they *cannot* get it to work quite right even in NYC: there are areas where you routinely wait for hours for a taxi.

            It’s frankly easier and more reliable to just run a train on a schedule. MASS transit is the correct solution when you have a MASS of people.

          • ROBwithaB

            It shouldn’t be too difficult to have a physical connection between individual minibuses, similar to those on trams and trains. Only problem is that it then becomes the opposite of a “crumple zone”. Unless a towbar extends telescopically… Hmmm…
            I avoid using my credit card for anything other than rental cars (and wish that I didn’t need a card for those either) and occasional overseas travel. But the truth is “they don’t need to know exactly WHO I am, only that passenger #543768 wants to go from A to B on most days at about 7:30 am and back again at about 6pm. With enough such data, they can get a great composite picture without needing any personal information. Which is as it should be.

            And I encourage people to :”steal” my ideas. I have now officially given up trying to commercialise every idea, because I’m running out of time. Starting a business takes an enormous amount of time and energy. I have about 50 different patentable inventions in various stages of development, and it’s clear to me that I’ll never be able to bring them all to market. I only have a few decades left.
            Anyway, secrecy is SO last century. Open source is the way to go. And it’s a lot more fun…

          • Bob_Wallace

            “And I encourage people to :”steal” my ideas. ”

            Good, because we’ve previously had this discussion here a time or two. Spontaneous carpool was a term I used in at least one of them.


        • Freddy D

          3km to the railway station – this is where protected bike lanes come in for the first and last mile. Come and go as you please at your schedule, short distance. And the top priority is feeding railway stations in a star-shape for 5km out. Less parking issue at the station too because they’re small and lockers / bike station rooms can hold thousands.

          • Carl Raymond S

            Did I mention the “very steep hill”? Perhaps an e-bike.

      • MarTams

        If you look at google maps and estimate the travel time between points A and B, including traffic conditions, travel by car is faster than public transit 99.99% of the time!

        In order to make mass transit better, whole cities should be redesigned around mass transit. Right now cities and suburbans are designed with the assumption that majority of the workforce will use personal vehicles. The city planning has been in collusion with big car manufacturers where mass transit was relegated for PR show.

        • Freddy D

          True when city density is low enough to get a car through and mass transit is poorly designed (bus-focused, “light rail” that goes down the middle of a street), as nearly every system built in the US is during the past 30+ years. Systems with dedicated rights of way, frequent service, and high vehicle speed beat auto gridlock in travel time without problem. Most well designed systems lie outside of the US, with a very few notable exceptions.

          To your point on designing cities, many folks under 40 (obviously not all) are favoring moving back to urban cores with pre-1950 layouts because they’re more pedestrian friendly and many find the aesthetic of the strip-mall / car-centric city or suburb ugly and / or inconvenient. Throughout the US, these parts of town are going through a renaissance and real estate prices have far surpassed suburban prices, reversing decades old trends. Developers like open fields because it’s easy development, but their customers are opting for something different, driving up prices of remodels. I don’t know what happened to the field of urban planning and architecture after about 1950 or 60, but it doesn’t seem to be standing the test of time.

          • neroden

            I don’t think you’ve actually been paying attention to recently constructed light rail systems in the US. Most of them are actually on old railroad routes…

            San Diego Trolley (1981): mostly railroad route, exclusive ROW
            Portland MAX (1986): mostly exclusive ROW near highways or on railroad routes, though there’s stupid stuff downtown
            LA Metro Rail (1990): nearly all exclusive ROW on old railroad routes, though there’s some stupid stuff in Long Beach
            St Louis Metrolink (1993): entirely exclusive ROW on old railroad routes
            Denver (1994): almost all exclusive ROW on old railroad routes or highway routes, except for a short section downtown
            Dallas (1996): almost all exclusive ROW
            Salt Lake City (1999): almost all exclusive ROW on old railroad routes
            HBLR (2000): almost all exclusive ROW on old railroad routes
            Seattle Link (2009): almost all exclusive ROW, even where it’s in a median
            Minneapolis (2004): nearly all exclusive ROW in medians
            …do I really need to go on?

    • Kraylin

      What a wonderful service that would be! You would still have issues with late people though… not sure what the right solution would be. Build in a few minute window and then just bail on the late person?…. Almost goes back to scheduling of Mass transit in general. I am still going to want my car because plans change and I am “late” quite often, even if only to my own schedule leaving a few minutes after I intended…

      • Bob_Wallace

        If you order a car for 5:35 you have x minutes to get in and close the door. If you reschedule far enough in advance there would be no charge. If you miss the ride, don’t reschedule, you pay for the seat you left empty.
        I don’t know what the time windows would be. That’s probably something which would need to be learned with use.

  • Steve Grinwis

    I think that’s pretty interesting, and thought provoking.

    I do think the autonomous car will destroy airline travel. I have no reason to fly, if my car can drive me there when I sleep…

    • Especially the Tesla model with free Superchargers. It’s already starting to happen with Tesla owners using Autopilot and Superchargers to drive much further than they would have considered previously. It’s going to be disruptive I think.

  • markogts

    If it is morally acceptable to reward car pooling, the same line of thinking should lead to a ban on empty driving cars.

    • Bob_Wallace

      That makes no sense. It will be necessary, at times, to move empty cars from place to place.

      • phineasjw

        Such as having your car pick you up at the airport.

        • Bob_Wallace

          Or having multiple self-driving cars take themselves to the airport to meet the last plane of the evening. There will be no passengers going to the airport.

    • Carl Raymond S

      How often does person A drop person B to their destination? The return journey with person A has the same utility as an empty car returning home, less utility in fact, as it prevent person A from doing something useful – the opportunity cost.
      A driverless taxi should spend less time empty, as it might pick up a ‘fare’ for the return journey.
      But I agree that empty cars are going to lead to plenty of road rage among those not looking at the bigger picture.

      • Freddy D

        Yep. What do I care whether my car drops me off at work and then just toodles around downtown creating gridlock all day long empty?

        • Carl Raymond S

          Fair comment – there will have to be guidelines in place to prevent active bastardry like that. Your car will have to continue through town and find some place less congested to spend the day. Better still, it picks up an Uber fare, makes you a dollar and performs a useful act.

          • Bob_Wallace

            You care (or should care) what your car does because it will cost you money. Best for you if your car takes itself to the nearest free/low cost parking spot. One with a charge outlet, if needed. Or operate as a taxi when you don’t need it and earn you some money.

  • Freddy D

    Thanks for the great article, and clearly thought-out set of rationale that better articulates what my intuition said: self-driving cars will increase car use, increase traffic, and increase congestion. I really don’t care if my car drops me off and just drives aimlessly around the block all day while I’m at work. For elderly, disabled, intoxicated, etc, why bother spending double the money to live in an urban setting and walk when I can get a cheap place further away and sleep my commute away. Urban sprawl 2.0?

  • JamesWimberley

    There is a simple and tested solution to traffic congestion: put a price on it. The economic logic is impeccable: putting a marginal car on heavily used urban streets slows everybody else down. It’s a standard externality, and externalities should be taxed.

    Ivory tower theorising that ignores political realities? It’s been introduced both in authoritarian Singapore and in democratic London, and has proved politically sustainable. Boris Johnson has kept it in London, and is SFIK under no pressure to stop it.

    • Marion Meads

      It’s being done in LA and to some extent along I-680 in the Bay Area, from Pleasanton to Milpitas… Still doesn’t help much during peak time.

      • JamesWimberley

        The price is too low.

    • Absolutely. Part of the point of bringing together this material was to point out that policy will need to adapt even more than it does today.

    • Freddy D

      True from a rational economic standpoint. In the US, anyway, road-welfare is a political sacred cow – everyone is “entitled” to the government providing, at tremendous expense and tax-breaks, free roadways. I’m not sure if politicians could pull this off, regardless of the sensibility.

      • JamesWimberley

        How does Boris Johnson get away with it in London? He’s a flamboyant populist Tory (imagine an intelligent Donald Trump), who succeeded ex-Trotskyist Ken Livingstone as Mayor. Red Ken brought the policy in, but Boris has stuck with it.

        • neroden

          No rural voters in London. No suburban voters either.

          That’s the basic political divide here. The *national* Tories tried to scrap the Congestion Charge. The *local urban* Tory kept it…

    • Bubba Nicholson

      A market solution would optimize travel speeds. If demand determined price on our roads on a constantly updated basis, our roads would never be clogged again. The programming is straight forward. GPS computers in vehicle tags would allow the government to keep track of every vehicle and charge what the market will bear or toll-by-plate toll collections fluctuating by demand. Ambulances, fire trucks, police cars, political dignitaries, and the uber-rich would be free to travel any road at any speed, provided they pay for the privilege of inconveniencing and/or endangering all the rest of us at the prices we each agree to pay for allowing that inconvenience and/or risk. Autonomous helicopter sales would likely improve as a direct consequence since overland travel costs would be valued appropriately. Street improvements and repairs could be similarly sequenced in economic rather than political order. This has been envisaged by various private companies for many decades now.
      Just don’t ask the government or the military to do it. Remember the long lines at military gas stations in Iraq and the “shortages” there? Like in the old Soviet Union, right? Aligning market forces to political fiat would be needed, presently something completely beyond the capacity of democratic government, but well within the corporate realm. We would have to bid out and sell the public roads to corporations who would behave economically, even sabotaging competing roadways to garner higher traffic and greater ROI. Such a system of more competition within the law would divert proper resources to travel needs appropriately. Roads would be far less expensive and everyone would be far happier.

      • nakedChimp

        somehow this doesn’t sound like progress:

        “by the 1790s in the German-speaking Holy Roman Empire in Central Europe, there were approximately 1800 customs barriers. Even within the Prussian state itself there were at the beginning of the 19th century over 67 local customs and tariffs with as many customs borders. To travel from Königsberg in East Prussia to Cologne, for example, a shipment was inspected and taxed 18 times.[2] Each customs inspection at each border slowed the shipment’s progress from source to destination and each assessment on the shipment reduced profit and increased the price of goods, dramatically stifling trade.

        Monopolies don’t belong in private hands.

  • Marion Meads

    Using autonomous vehicles can boost productivity and recreation, reduces stress too. You can do your computer related work while your autonomous driver is doing an average 2.5 mph on the road. Ha ha ha! Being productive during rush hours instead of being stressed would be the biggest contribution of autonomous driving and it would be independent of traffic congestion.

    • Freddy D

      And instead of taking mass transit and working on the train, I can take my car in work in it. And we’ll all do it and then what happens to congestion?

      • MarTams

        Real experience with BART from San Francisco Airport to our house in Union City took 3-4 hours several times already. If we drove, would have taken 1.5 hrs in the worst traffic and only 25 minutes during off peak hours. Public transit is very slow door to door and would be slower still if everyone uses it. Part of the problem is that you’ll have to wait for layovers as you transfer from different lines or change trains and busses at each hub. There are no direct routes for majority. While I appreciate the work out for the last mile to our door, that last mile of walking adds another 30 minutes.

        • Bob_Wallace

          You’re making a bad comparison there. From Union City to SFO by BART one has to go north and around the Bay. By car you cut across the bridge and make a much more direct trip.

          You’re saying nothing about how well public transportation works by picking a place where there is no public transportation. Were there a BART route from Fremont to Millbrae/SFO then the trip would be a zip.

          From Union City BART to your house would be the self-driving car task.

    • Bob_Wallace

      Yes. Letting the car do the driving frees up the driver to do something productive and/or enjoyable during the commute.

      Autonomously driven cars should reduce accidents to almost zero. They should allow smooth lane changes. They should optimize routes, reducing the number of cars on a particular road somewhat. All that will help but self-driving cars will not solve all the problems. If people can’t be enticed to move to mass transit or adjust their travel time then at least the commute wouldn’t be tortuous.

    • Kraylin

      Agreed. Even if I still had to endure the stop and go traffic for 1-1.5hrs on the way home at least with an Autonomous EV it would be quiet and as Marion said I can focus on something else. Taking the train into downtown Toronto from just outside works well but it takes just as long and now I am subjected to the crowds, scheduling, etc of the train… The trains here are pretty decent really but I would still much rather have my car, but the stop and go traffic is a stressful waste of time…

    • MarTams

      Entrepreneurship idea: Build a commuting flying autonomous drone service for congested cities. Autonomous flying drone technology is almost here. Simply make it more efficient and increase payload to carry one or two passengers. The drones can communicate with each other to avoid collisions. It can be made electric and automatically plug themselves in for a recharge. Imagine all the time you will save and the various toll fees. No need to build new roads and no more congested roads. People will pay a large premium for this service. And when not a lot of people use them at off-peak times, these drones can be used to deliver goods. There is no idle time. The potential is so huge!!!

  • eveee

    Lets get to the meat of the matter. Roads. There are some improvements in throughput by various means, but in the end, roads can only carry so much traffic in so much time.
    Conversion of roads to transit makes more sense than autonomous vehicles. So I agree, congestion is not a dominant reason to use autonomous features. I am not so sanguine about truly driverless vehicles for various reasons.
    On the scheduling or path issue, if you think about it, it doesn’t change much in some situations. The demand during peak times won’t change just because of autonomous cars. Autonomous cars can smooth demand when there is no hurry for a car to arrive, and the car is scheduled in advance, but it can’t do anything about the rider that didn’t plan and wants it right now, during rush hour.
    So I don’t see how autonomous is going to substantially improve rush hour traffic.
    Telecommute has more impact, IMO.
    As long as there are daily 9 to 5 jobs, there will be rush hour traffic. Frankly, its more basic than that. Circadian rhythms.

    • Freddy D

      Precisely – roads and automobiles do not scale with increased density. All large and medium US metro areas have long hit saturation, where widening a lane here or there is a hopeless strategy in the face of another 20% population increase. What transportation can scale? Well designed railroads can scale massively – 30,000 per hour per lane vs 1,200 for freeways. Bicycles scale in the urban core very nicely, carrying far more than a lane of cars due to their compactness – see New York City’s protected bike lanes and the traffic studies before and after.

      • eveee

        Hey thats great input. You have any links to any of those studies or articles? I would like read up.

    • Bob_Wallace

      If there were autonomous cars within dense urban areas people might be willing to use mass transit to move from suburban parking lots into the city.

      Mass transit generally needs to be more comfortable. In particular, less like human sardine cans during rush hours.

      • Frank

        I agree, and I further think that autonomous driving could lead to more car pooling. There is no reason a car couldn’t be waiting at the train station when the people arrive, and pick up 4 people heading in the same direction, as long as the computers work out the details. If your ride cost half as much, wouldn’t you?

        • Bob_Wallace

          Four people? 25% as much.

          Not having to deal with traffic and parking? I’m in.

          You don’t even have to tip the cab driver.

          • Or six people, or …, Wait now we are describing a transit system. Oh yeah, if cars can be autonomous then so can transit, buses etc.

            Of course if the cars are really autonomous then won’t they refuse to drive into congestion. “Hey friend, I’m not taking you there, we’re going to a transit stop. When you get close get one of my buddies to take you the final mile or less.”

            Then again, autonomous transit has existed for decades. They are called elevators.

    • MarTams

      Why not concentrate all of our efforts on designing more efficient autonomous commuting drones that fly? No need for roads! The ultimate solution of road congestion! Flying drone technology is almost mature. Just need to increase their load and efficiency, and yes they can be electric and programmed to recharge themselves too!!!

  • Matt

    “The top 30 bottlenecks are each responsible for more than one million hours of lost time annually.” Makes you wonder about those 30 spots, are they all cases where light rail would address the issue better than more roads.

    • Freddy D

      Great point. One minor comment – Light rail over promises and under-delivers on people-per-hour per lane. Most US implementations don’t do much better than a lane of cars because they don’t have a separated right-of-way. Separate the trains in their own tunnel / corridor, and then you really have a solution proven to work. Also, size the trains and stations to meet the demand – another common failure of US urban passenger trains over the past 30 years.

      • Kraylin

        We (Ontario and City of Toronto government) are in the process of building a major subway/lrt above/below ground split system across the Middle of Toronto. (Named Crosstown). Our first new line in a very long time.

        I wonder if the stations are being built for the very long term. I hope they are building them so big that at first they will seem empty to allow for major growth in the distant future. Going back and expanding these things is so expensive they never do it and you end up with an over crowded system seemingly impossible to fix….

        • Freddy D

          “Allow for major growth” – easy to answer by the following questions:
          1) Does the route have 100% of its own right of way? (no sharing with cars, pedestrians, etc.)
          2) do the platforms and route accommodate 10 car trains?
          3) can the trains run 2 minutes apart?
          4) will the trains run at highway speeds? (130k/h or 70mi/h)

          If yes, then the system can move real volumes of people. Most of the US systems over the last 30 years fail on some or all of these attributes and end up carrying 700 people per hour rather than 30,000 like the real systems. Toronto is one of north america’s biggest metropolises. It would fill up the 30,000 person per hour system in a centralized route. Perhaps a dozen such well placed routes. LA would fill a couple dozen routes like that (wow, traffic in LA might be bearable then…)

          • neroden

            Short platforms are a recurring problem with recently-built systems, yes.

            A three-car train still holds six times as many people as a bus, so if you can run them every 2 minutes (which you can on every system I’ve ever looked at) it’s a lot of capacity.

        • Truth hurts

          The Toronto subway is the worst ever. It is slow, the layout is a joke. I dont know what was in the heads of the people building it. If european cities could not serve as an example for some reasons, the should have taken a look at the Montreal subway at least.

      • neroden

        Most US implementations of light rail do have a separated right-of-way, so I have no idea what you’re talking about there, but I think you simply haven’t done your research. Yes, the ones which share track with automobiles get delayed horribly.

        Grade crossings (with crossing gates, flashing lights, etc.) are usually OK and save huge amounts of money over bridges, though there are still idiots who drive around the gates. You can’t fix stupid.

    • Doug

      There are a number of good ways to fix the problem of congested bottlenecks that will allow efficient traffic patterns with and without autonomous vehicles. It just takes determined effort and courage.

  • MarTams

    We should develop a system of scheduling the use of main roads so that it will not be congested. If you did not logged in at the designated time, those in waiting list can take your place and you go to the waiting list. Those who don’t have schedule should queue in the waiting list line. I’d rather do something else and wait for my turn rather than inch myself in a crowded freeway.

    • eveee

      There is a waiting list. Sitting in traffic, LOL.

      • Marion Meads

        You could wait at your place of work until your allotted time comes up. Unlike in the picture above, which is typical of bumper to bumper traffic all over major cities, they are burning gas, and lots of it, for nothing. At least EV’s could dramatically reduce such types of emissions. Rather than be in that freeway parking lot, it would have been better to wait for your turn while doing something else, such as work or recreation, as long as it is guaranteed that the roads are not very congested when your turn comes up.

        • eveee

          Really? You want to sign up for waiting after 5:30 when work is over or do you want to go home. Judging from that picture of traffic, most people can’t wait to go home.
          All the more reason telecommuting is a much better alternative than autonomous cars.

          • Carl Raymond S

            Marion is right. It’s the same as the supermarket delicatessen. As some, you wait in a queue. At some, you take a number. While you wait for your number you can buy your cheese. I’d rather take a number and buy my cheese.
            Agree, telecommuting is the fastest – only trouble is, it’s too easy to walk to the fridge and eat said cheese.

      • MarTams

        When I see something like that, I wished I could fly over the traffic. So why not autonomous flying drones that can carry you from point A to point B? For sure road congestion would be a thing of the past and you bypass all those exorbitant toll fees. Imagine that! No need for new roads. Each drone communicate with each other so there is no midair collison.

        • ROBwithaB

          Did you see the new Chinese passenger drone featured at CES? Pretty much what you’ve envisioned. It already exists.

          • MarTams

            I am well aware of it. That’s not even a pre-alpha version. It is still very buggy and inefficient. It needs to have turbo prop and safeguards to prevent it from becoming a meat grinder. The Chinese drone doesn’t even work reliably yet. So we need to refine the technology.

        • nakedChimp

          So you want sky-congestion?
          With every Bob and Betty zipping around on the sky?
          The fastest and loudest will be the coolest for sure?
          Those ‘rare’ fixed wing aircraft dump enough noise into the environment already (pretty cool as you reach more receptors from high above), there is really no need to increase this by multiplying their numbers with 1000 or more.

        • ROBwithaB

          Because physics.
          Staying aloft by means of rotors is very energy inefficient. And you get punished heavily for any increase in mass, much more so than with a car, for instance.

          Range is always going to be very limited. Failure is likely to be very fatal.

Back to Top ↑