3 ms·

Road cost efficiency increases the more small vehicle traffic you put on it. Think high speed autonomous caravans. The cost per mile is already too low. That’s

• shkkmo 8y ago
  > Road cost efficiency increases the more small vehicle traffic you put on it. Think high speed autonomous caravans. Higher density vehicles (like buses) increase road cost efficiency even more. For long distances, rail is far more efficient in terms of both energy and infrastructure cost. > The cost per mile is already too low. That’s the over-demand problem being discussed in TFA. The article doesn't really have a clear thesis and its (unsupported) title sort of contradicts that. The best I can tell is that the articles tries to argue that single-occupancy car trips should (and will) become more expensive because of no longer being subsidized. > Autonomous driving technology will triple existing highway capacity by ~2050 when it becomes illegal to drive non-autonomously. That’s a lot of “free” infrastructure. The wasted space of, and maintenance costs for, that infrastructure makes it far from "free". In the short term, the degree to which autonomous cars and ride hailing make single occupancy car trips more available and affordable is going to increase the congestion and amount of needed infrastructure. > We should not invest billions in light rail that will be unprofitable its entire life and archaic in 50 years. The trade-offs of light rail vs. buses is complicated. I highly suspect that buses, and infrastructure to support buses is often the better choice. That is especially true given the flexibility we need in the face of a changing transportation landscape. That doesn't justify "self-driving cars" mean mass transit "makes no sense". For some of that mass transit, rail is still the best option.
  • zaroth 8y ago
    My point about autonomous driving software increasing highway capacity granting us "free" infrastructure is not to say the underlying infrastructure is free (hence the quotes). But when existing infrastructure becomes 3x more efficient due to private market software development, you have effectively tripled the value of your infrastructure for free. An inner-city bus ride costs $2.00 and doesn't get me door to door. The cost of getting to the bus and getting from the bus stop to my actual destination pushes that, even at minimum wage, probably closer to $5+. That's 30 miles (at cost) door to door in a hypothetical Tesla RoboTaxi! How can a bus compete with that? Yes, a bus is a much more efficient use of space on the road, but it is not more efficient to the individual riders. Hence the negative externality. But in reality, the real-world usage (market-share) of buses and light rail is so negligible as to be almost entirely irrelevant. "US Passenger Miles" according to the Bureau of Transportation in 2017 for Highway travel is 5,502,417. That number is in Millions. Whoa... That does include 365,220 labeled "Bus" which appears to be private charter, not public transit. In the "Transit" section "Bus" is 19,343 and "Light rail" is 2,795, "Commuter Rail" is 12,321 million passenger miles. That means Bus, Light, and Commuter Rail total 34,459 million passenger miles, which is ~0.6% the number of Passenger-Miles traveled by private vehicles on the highway. By the way, for longer rail trips, Amtrak is 6,527 million passenger miles in 2017 (declining year over year) and air travel is 693,818 million passenger miles and growing year over year. Heck, lets get electrified air travel and ditch trains for passenger travel as well. Heavy freight via train will still possibly make sense, although autonomous trucking will probably steal a lot of market share. For hazardous materials and national security purposes you probably still need the heavy freight network to be operational. It would be cooler if the really big stuff all went via dirigible though. [1] - https://www.bts.gov/content/us-passenger-miles https://www.bts.gov/content/us-passenger-miles
    • shkkmo 8y ago
      > My point about autonomous driving software increasing highway capacity granting us "free" infrastructure is not to say the underlying infrastructure is free (hence the quotes). But when existing infrastructure becomes 3x more efficient due to private market software development, you have effectively tripled the value of your infrastructure for free. The alternate way of looking at is that you will be paying 3x more in maintenance and space usage then you need to. By not planning ahead for this "certainty" you have tripled the amount you need to spend on transportation infrastructure in the future. In the intervening 30 years, you will have had to deal with spiking demand from how self-driving cars reduce the cost of car trips. To prevent having to build a bunch of infrastructure that will eventually become unneeded, mass transit is critical to avoid wasting money and space. We need to ween ourselves off of needing all of this road capacity with usage and congestion based surge pricing. As self driving cars begin to increase usage efficiency, we can slowly lower the usage fees charged to self driving car to further incentivize the transition to self driving cars. > An inner-city bus ride costs $2.00 and doesn't get me door to door. The cost of getting to the bus and getting from the bus stop to my actual destination pushes that, even at minimum wage, probably closer to $5+. How can a bus compete with that? By charging a usage fee to push riders to more space and energy efficient forms of transit. There are also other improvements that can be made to improve last-mile transit options. My skateboard does a very good job of this and I expect to see more scooters. I also think that autonomous vehicles have a large role to play in helping with these last-mile trips in less dense areas where the "last-mile" may be several miles or more and space is at less of a premium. > the real-world usage (market-share) of buses and light rail is so negligible as to be almost entirely irrelevant. The good thing is that this give us a lot of space to compensate for the increased demand for transit that self-driving cars will bring.
      • zaroth 8y ago
        Whether you look at autonomy as granting the ability to reduce your future infrastructure development costs, or providing a boost in road capacity without increasing infrastructure cost, either way autonomy is providing a massive boost / making infrastructure more efficient and thereby effectively lowering (possibly halving or more) roadway cost per passenger mile. Why would the road infrastructure ever become unneeded? It's more needed than ever with the rise of incredibly cheap, extremely convenient, on-demand private transportation. The availability of which increases productivity and drives overall GDP growth and prosperity. I don't see why it follows that we need to ween ourselves off of roads. Roads are the backbone of our economy, and more efficient use of roads provides huge economic benefit. More availability of on-demand low cost private transport is an amazing benefit to low-income workers. Usage fees for roads that are being used more efficiently than ever, to try to drive people to less efficient (for them) modes of transportation, is a recipe for suppressing growth. It's a regressive tax which disproportionately impacts the poor. Scooters are neat. Scooters are "0.1" relative to cars "5,500,000" in the 2017 million-passenger miles table. Maybe that grows 1000x to be a "100". But in a fully autonomous future, scooters are obsolete and not worth the investment in designated travel lanes. Creative future designs for super-light autonomous personal transport will certainly not take the form of a "scooter".