Driverless Congestion
ethz.ch
ethz.ch
One of these levers is to quadruple road capacity through new roads.
The other is to quadruple road capacity through driverless cars with perfect routing efficiency.
What do these have in common?
First: they are still massively wasteful of space. Moreso with a driverless car, which is going to be able to avoid having to find a parking spot nearby.
Second: induced demand. Double capacity, triple, quadruple, it doesn't matter. It will return to the "barely works" equilibrium in short order, except you'll have further lashed yourself to a wasteful mode of transport.
I recognise cars have unique utility and that driverless cars will enhance that utility. But for mass people movement, a train is the clear leader, followed by a bus in a dedicated lane. As the world becomes increasingly urbanised that will only grow more important.
It's harder than just providing parking space for the cars and there are problems with that.
Whether we increase the number of cars on the road by packing them tighter or adding roads doesn't matter. Any gains are quickly consumed.
Only if wasting energy is acceptable. Loafing about doing loops around the owner will very quickly be more costly in terms of energy expenditure (plus opportunity cost of "now I have less energy to go to my next destination") than parking for awhile, especially if we get to a point where every parking space is capable of EV charging (which is exactly the direction in which I'd like to see the world gravitate, since that'll significantly alleviate the problems with EV range relative to ICE).
Thus, I'm far from convinced that this will ever actually be a significant problem. Only way the "perpetually-circling autonomous vehicle" problem would be realistic is if we lived in some Nikola-Tesla-esque utopia where wireless electricity is abundant and constantly charging cars as they drive (or maybe if it's a sunny day and solar tech gets advanced enough to be able to meaningfully power a car on solar power alone).
So to recap: the car is used twice as much and yes, does laps of the block while waiting to pick you up.
The ideal will always be for it to park (or better yet: charge) as soon and as close as possible, and to delay "unparking" and departing for pickup until as late as possible. It's possible that some people will want to subvert that somewhat (i.e. "please park at this specific parking garage instead of the closest" or "please loiter for a couple minutes and wait for me"), but - again - I ain't convinced those will actually amount to a significant traffic impact.
EDIT: as another bonus, self-driving cars don't have to worry about whether or not the driver can get in/out during these "wait for my owner to need me again" periods, so they can park in much tighter parking spaces while they wait.
Ahem... a 7 minute ride from our hotel to the train station came up to 35 CHF. That's around 5 km, so around 7 CHF/km or 6.25 EUR/km. Just something to these dreamy 56 rappen and realism of the study in perspective.
If you look at the full context, I think the article states that this is the price which could be achieved by an optimally sized fleet of driverless taxis - it even says right there that the actual, current price with conventional taxis is much higher than that figure.
The current taxi fees in Zurich are actually even higher than 2.73, even before factoring in the base fee and fees for waiting times, but I don't see how that in itself would limit what prices driverless taxis can realistically achieve.
[1] https://www.stadt-zuerich.ch/content/dam/stzh/pd/Deutsch/Sta... [2] https://www.uber.com/de-CH/blog/zurich/preise-zurich/
First, we always overestimate new technology’s ability to solve problems, while underestimating the side effects. So any grandiose claims about self driving cars fixing congestion should’ve been met by more skepticism in the press.
Second, we already know that people are willing to spend a large amount of money on personal transit that goes above and beyond the bare necessities of transit, at least in rich western nations. I see no reason why that would suddenly stop once the vehicle’s owner is no longer behind the wheel.
The success of Uber suggests that some people are choosing not to purchase private vehicles. But we don't know if this is limited to Millennials.
The success of Uber only tells you that a) some people use taxis, and b) a pirate taxi using VC money to offer rates at below market level will be popular with taxi users, for obvious reasons.
E.g. the minicab and Uber drivers around mine are often more clueless than I am about traffic patterns nearby, despite the fact I don't drive myself. Realtime coordination to pick less congested routes dynamically could improve things a lot.
And apps like Citymapper that don't just suggest a single mode of transport has a lot of potential. In London they're experimenting with bus routes of their own. The long term potential is being able to promise you short waits and low cost by e.g. telling you to wait for the bus when the bus is near, but sending a car - rideshare or not - if there's a wait, and dynamically scheduling minibuses for high traffic routes not covered by the standard routes, for example.
I'd prefer that over a car the whole way a lot of the time, because living somewhere highly urban, driving the whole way can almost never compete on speed, but figuring out the optimal route that keeps changes few and waits low is often tricky. E.g. Citymapper just recently introduced me to a route I had no idea was an option - a station I've never used before because it's too far too walk and awkward to get to by bus, and so I didn't even know what trains go from there. But suggesting an Uber there and train from there, and a second change got me where I wanted to go ~20 minutes faster with the same number of changes as if I'd gone the route I expected to take. Mixing and matching like that has the potential to make taking a car the whole way a lot less attractive for a lot of routes where people do opt for that today.
But it does require a lot of things to go right. For starters, multi-mode transport apps that are free to pick the optimal mix needs to win the battle for being consumers first choice in planning trips.
Ridesharing is a greenwash smokescreen for an industry that predicts yearly driven miles to go up by 300%.
If you need it for a daily drive (in a low demand route) buying it is probably cheaper, if not, just pay per use.
People who used to buy DVDs now use Netflix. Companies who used to buy serves use AWS.
Owning some things do make sense though
https://www.marketplace.org/2017/08/15/business/one-vintage-...
For people without a hired person to manage all these subscriptions, I wonder just how cognitively-taxing a rent/subscribe-everything lifestyle would be.
Not sure where you're from, but for me and my two vehicles, it's not really that big of a deal.
* Oil - look at sticker on windshield (when I change my own oil, I just erase the sticker and write it myself)
* Water - look in the radiator every once in a while; low? Add some coolant (never plain water!)
* Tires - look at them every now and again; need air? - add air, tread look too worn? - buy new ones, hear a clicking while driving? stop, and look for rock or nail
* Brakes - look and feel, listen for squeaking; if it squeaks, probably hitting the indicator strip - change 'em out. If feeling spongy, flush and refill fluid; pull the tires off (during a rotation) and look at 'em. Rear drums if you have 'em - common on trucks - only need to be checked occasionally - they rarely need to be replaced often unless you haul loads regularly
* Taxes - what taxes? I buy used, pay with cash - anything else, you're getting ripped off
* Insurance - get the bill, pay the bill
I do as much maintenance as I can myself, otherwise if I think it needs to go into the shop - into the shop it goes. Having two vehicles means I don't have to worry about if one is in the shop. Sometimes I've swapped - one car in the shop, get it back, swap the other one in (usually for things I can't handle - things that need a lift or are beyond my skill or tool level). About the only other thing I have to worry about is timing belts, but that's only on one vehicle now (my Jeep has a chain, I'll hear the noise once it wears beyond a certain point - the 4.0 I6 is near indestructible anyhow), and I got it changed when I purchased it, so it's good for 60k, at which point I'll probably need a complete engine overhaul anyhow (given the age of the vehicle - 1999 Isuzu VehiCROSS).
I love both of my vehicles, and I can't yet get anything in a hybrid or electric that compares (nor could I afford it, even if it did exist). I sincerely hope that battery tech gains to the point where a charge will get you 600-800 miles (because once you put that tech into a 4wd vehicle, with a lift and 30+ inch tires - that number will drop like a rock). But it ain't there yet, and even when it does happen, I won't be able to afford v1.0 - so I'm sticking with my current vehicles.
Plus - I'm hoping my VehiCROSS becomes a collectible; it has all the categories of one, but for some reason, it hasn't quite happened yet. I love driving that beast, tho...
I like being able to open up Citymapper, tell it where I want to go, and now they've started not just suggesting Uber and others for the whole ride as an alternative to public transport, but suggesting it for parts of the way. A car you own can't compete with taking you to a train station and picking you up on the other side.
Today it takes ~3 minutes on average to get a car to the front of my house. That's with human drivers, spaced out inefficiently. Get driverless, and you can get a mix of ride shares and individual cars based on spacing out cars based on predicted demand and picking the most efficient solution based on customer preferences with very low wait times at even lower cost.
Owning a car just can't compete in my eyes. Maybe that's shaped by living in European cities where driving long distances almost never competes favorably with mixing and matching with trains and underground (a typical commute for me in the days I used to commute in to central London would have taken twice as long with car, and maybe ~30% less if a car/rideshare service could guarantee to be here in 3-4 minutes instead of waiting for the bus or walking to the train station; with apps suggesting the best route I'd still take the bus when the wait is low, but know I could start my journey later because I'd have a shorter maximum wait)
EDIT: Even more so because parking restrictions are getting stricter and stricter. The London borough where I live won't even entertain planning applications for housing if it includes more than 1.5 parking space per unit. That's expected to drop further as density increases - note that this is not permits for the road etc., but applies whether or not you build your own off street parking. The reason is that restricting parking is seen as a way of discouraging car ownership to reduce demand on surrounding road infrastructure.
But if you can afford a car, it's still a big help. So families will keep buying cars. I'm just trying to explain GPs "I don't understand why people buy cars", not saying that it's a requirement.
Having lived almost my entire life in 1M city, I don't have one either, but I recently moved to a ~15k city and this is forcing me to finally get that license.
It's possible the math would change if I both lived and worked within walking distance of a light-rail stop, but only a small percentage of residences or offices fall into that category, and neither my home nor my office are included.
It probably also helps that my car is paid for, so my monthly costs are only: insurance, fuel, 1/12th of my inspection and registration fees, and a few oil changes per year. Over a long enough time frame, I'd have to include whatever it will cost me to replace this car someday, but eventually the VC money underwriting Lyft is going to run out, so I'll take my chances.
So, what are reasons not to own a car in the age of self-driving cars?
- it could be cheaper for a lot of people
- you'd be more flexible: use a small car to commute, a big car for shopping etc.
- there would be enough self-driving shared cars available so you never have to wait a long time to arrive
- it saves a lot of resources if less cars have to be built and their utilization is higher
- you can drive new cars as soon as they are added to the pool of shared cars
- you don't have to deal with car maintenance
- you are more likely to use your bicycle / ebike
Here are some reasons against using shared cars:
- mass surveillance - monitoring the movement of almost every citizen
- shared cars may not be as clean as your own
- can't store items in shared cars (baby seats!)
- favourite car may not be available all the time
- there may not be enough shared cars avaialble in your area, especially if you don't live in a city
- you have special requirements such as baby seats, 4WD or you want to drive a particular car model all the time
- prestige
- Being able to get where you need to go without requesting a ride. - Your own car will wait with infinite patience. - Kids
If your car has parked far away from your location you need to "request" it just like a shared car.
> - Your own car will wait with infinite patience.
That could be useful every now and then. But not very important on my list if the waiting time is small enough.
> - Kids
I mentioned baby seats. But the availability of shared cars could mean that you can drop off multiple kids at multiple locations using multiple cars all at once.
Sure, but you'd be the only one requesting it, as opposed to a bunch of other people requesting it, too.