Peak Traffic
Reader JT says that self-piloted cars that dropped you off at work and then went home “would double our current traffic problems.”
Maybe yes maybe no. I don’t want to get too deep into the issue of what could be done with hypothetical robot cars. But it’s important to recall that our “traffic problems” aren’t really a problem of total road capacity, they’re a problem of peak road capacity. I once left DC at 11:30 PM on a weeknight and drove to Richmond. You’ve never seen I-95 so clear of traffic. And you’ve probably never seen it for the same reason there was so little traffic — nobody drives at that hour. Which is part of the reason why a well-designed congestion pricing scheme can do an enormous amount of good. Relatively small financial incentives aren’t going to compel people to radically reorganize their lives, but they very much can cause people to shift low-priority trips out of peak usage times. That leaves people who really do need to travel at peak times stuck to pay the fee, but in exchange for their fee they get shorter, less annoying commutes. Right now we have roads that are very frequently under-utilized at the same time that most of our trips occur during periods when they’re overtaxed. More transit, etc., is part of the solution to that problem but so is simply spreading the driving around so that somewhat more of it occurs at low-usage periods.
39 Responses to “Peak Traffic”
Register
RSSLogin
Topic Cloud
ArchivesBlog RollAbout Matt YglesiasContact Me
Donate
October 14th, 2008 at 12:10 pm
This experiment has already been performed, re: Kei-Cars in Japan. Dramatically smaller cars and tough rules on parking spots can have strong effects on traffic.
October 14th, 2008 at 12:11 pm
My only thought about congestion pricing is this: Won’t you create a class system with regards to the hours you work. Those who can afford it, will work normal business hours and get to be home with their family and kids at night. Those who can’t will leave before dawn or get home late at night.
October 14th, 2008 at 12:15 pm
The best way to cut down on peak period congestion is to encourage more telecommuting. Unfoirtunately, the resistance from government and corporate managers has frustrated this approacy. For instance, Congressman Frank Wolf of Virginia has spent much of his Congressional career banging his head against the wall trying to encourage government offices to make more use of telecommuting to little avail.
October 14th, 2008 at 12:27 pm
why don’t we view the congestion and accompanying annoyance and wasted time as already priced into drivers’ decisions? surely if drivers, as at least some must be, are already aware of the traffic conditions in or into a major city at peak hours, they’ve already decided to bear some nontrivial cost; in this case, inching ever closer to a driving induced heart attack. what’s the problem then? poor information? stubbornness? stupidity? or what?
October 14th, 2008 at 12:34 pm
I agree with SLC. Telecommuting can have significant impacts immediately without spending much money – when compared to other approaches (mass transit, walkable neighborhoods, even bike lanes) the impact is much larger, much faster and much less costly. It can be phased in – even allowing someone to work 4 hrs a week at home, where one day a week they could come in at 10 and leave at 2:30 has a huge impact(when compared to other approaches). I can speak from experience having telecommunted for years. Many jobs can be done entirely remotely, and it is easy to expand this pool of jobs with better telecommuting tools, (conferencing, Instant messaging, presence (so you know if you co-workers are available…). Telecommuting is effective today and can be made much more powerful with modest additional investment
October 14th, 2008 at 12:56 pm
One of the most notable effects might be a reduction in greenhouse emissions.
The average vehicle sold today accelerates far more quickly than those sold 30 years ago. A V6 powered family sedan sold today has both more horsepower and a lower 0-60 time than most sportscars prior to about 1985. The appeal of this performance is purely in the pleasure of driving; fast and slow cars arrive at their destination nearly simultaneously.
If humans are no longer driving there is little reason to purchase a vehicle with a more powerful engine. I suspect most passengers will be so engrossed in conversation, TV, a book, or what-have-you that engine performance will quickly disappear as a major selling point. Smaller engines consume less fuel and therefore emit fewer greenhouse gases.
October 14th, 2008 at 12:56 pm
why don’t we view the congestion and accompanying annoyance and wasted time as already priced into drivers’ decisions?
Nick Paumgarten from the New Yorker said it better than I could: “People may endure miserable commutes out of an inability to weigh their general well-being against quantifiable material gains.” (Link to article here)
I can personally attest to this… near the end of the dotcom boom I got a job at a networking equipment company in the northwest suburbs of Chicago. It was 45 miles from home, but I figured that the money (it was $15,000 over what I’d been making before) was worth it, and besides, I could move if I found the commute intolerable.
It turned out that most of the additional money went to gas and auto maintenance, and I spent so much time in the car that I had very little time to spend with family and friends. In the general quality of life sense it was a mistake to take the job, but it seemed attractive at the start.
October 14th, 2008 at 1:07 pm
As I said above, I think it’s the collision of “you have to be at work from 9AM-5PM” and “family dinners are important”. Plus, the cost of moving is certainly nontrivial (dual careers/commutes, kids in school, selling/buying a house). Sometimes there’s very little choice.
October 14th, 2008 at 1:17 pm
capacity, or lack thereof, is not the sole factor in causing traffic. the irrational behavior of the human mind (changing lanes to find the fastest lane, slowing to view an accident of stalled card, impatient accelerating followed by braking, etc.) is also to blame. in fact, when capacity is challenged, this behavior both worsens and it’s effect becomes more evident. in principle, “robot cars” would circumvent this problem. thus, at a given traffic density, the probability of congestion would be decreased as compared to human drivers.
October 14th, 2008 at 1:20 pm
1. Agree with other commenters that it’s already more expensive (in terms of additional time and stress) to travel during peak times; why would a fee change anything?
2. Regarding telecommuting: in addition to managers not liking it (they’re afraid their employees won’t really work), employees are cool on it as well because if they can do their work remotely from home, somebody can do the same work remotely from India for cheaper.
October 14th, 2008 at 1:33 pm
With congestion pricing, the wealthy will rule the day. The poor will disappear into the night.
October 14th, 2008 at 1:34 pm
Matt’s made an illegitimate move in the argument: he’s simply changed the subject. Would robot cars add more hours on the road? Yeah, but let’s talk about shifting driving to off hours. What does that have to do with robots? Or maybe since they’re out of scifi movies, they have a natural predilection for spookier backdrops.
Anyway, congestion is only a secondary problem compared to getting people and cars off the road, period. And the whole thing shows again a 20-something male’s weakness for futurism. Call me when it happens.
October 14th, 2008 at 1:45 pm
Assuming self-driving cars with a good AI, they could also drive much faster with much less clearance than human drivers are capable of doing safely (or even unsafely). You could cram a lot more cars onto a highway and keep them moving at an energy-efficient 55 mph if an AI was doing it rather than thousands of independent human drivers with no knowledge of conditions up the road and limited knowledge about the intentions of other cars, all while generally behaving irrationally.
Another note on the return trip thing; this would largely eliminate the need for most families to own multiple cars. The car could take someone to work, drive home, someone else could use it for grocery shopping, and then go to pick that first person up from work at the end of the day.
We get even more efficiency if we assume people are using the taxi model – then we can minimize the time the car isn’t carrying a passenger, which is the most efficient of all.
It should be noted that what we’re really talking about at this point though isn’t too much different from Personal Rapid Transit, which is an idea that’s been around for quite a while now.
October 14th, 2008 at 1:45 pm
Responding to wahoofive, behavior would be changed with a very high fee. Once enough traffic forced into off-hours, the expensive hours become less crowded and worth paying for, for those remaining that can afford it.
Let everyone else eat cake in the dark.
October 14th, 2008 at 2:08 pm
I, for one, welcome our new robot car overlords.
October 14th, 2008 at 2:10 pm
While basically true, one problem is that (here in California at least) those times when “nobody drives” are the times when Caltrans decides to do major road work, shutting down multiple lanes of freeways and bridges. Sometimes I have gotten stuck in stop and go traffic trying to cross the Bay Bridge at 1 am.
October 14th, 2008 at 2:30 pm
A couple of points:
It seems less likely to me that owners would send their cars home than that they’d send them somewhere close at hand but farther than would be convenient to walk. This would reduce the number of people who drive around looking for parking spaces in dense areas. In addition, you can pack more robot cars in a given lot than human-controlled cars because there is, for example, no need to allow space for people to enter and exit. This also mitigates the risk of vehicle break-ins in unattended lots.
More importantly, I think people will feel less need to own cars they no longer personally drive. The worst part about getting a taxi is that paying a stranger for a ride is an uncomfortable situation. I think an automated taxi service would be wildly popular and cheaper than taxis are now. Routing these taxis would also be super-efficient.
I, for one, welcome our robot-car overlords.
October 14th, 2008 at 2:52 pm
There are freeways in parts of California (the East Bay and Santa Rosa come to mind) where,a side from driving in the middle of the night, there is no significant difference between peak road capacity and total road capacity. While there are times when the freeways are all but undriveable (insert a “nobody drives there anymore, the freeways are too crowded” joke here), most of the time they are at least awfu.
October 14th, 2008 at 3:19 pm
Auto-autos would mesh well with mass transit hubs too. No more need to have massive parking lots next to the station – cars could go off and service the local community, or park and recharge in some spot less centrally located. (& pickup at the other end too.)
Autoautos need some gnarly software though – it’s not quite as easy to synchronize the speeds and reactions of a train of cars as you may think. How does the car at the back learn it needs to break like crazy when the car in the front detects an emergency? A friend working on various auto-driving projects at UCBerkely told me about other problems as well.
October 14th, 2008 at 4:00 pm
This is just a note to say that, as another mid-20s white male, I wholeheartedly share any such alleged futurism fetish.
October 14th, 2008 at 4:02 pm
I’m wondering how far a leap from heavy congestion pricing to total ban on private automobile in city centers (Transport vehicles would of course still be allowed, w/ fee.)
With adequate transit and out-lots, seems perfectly reasonable to me…
October 14th, 2008 at 5:31 pm
Re: Which is part of the reason why a well-designed congestion pricing scheme can do an enormous amount of good.
congestion periods are periods when people don’t have a choice about travelling, or they would already avoid doing so. Very few people enjoy traffic jams after all.
October 14th, 2008 at 6:14 pm
I strongly agree with SLC that telecommuting is the single best option of reducing traffic.
This automatic car thing is nonsensical. More thoughts at: http://lowtechtimes.com/2008/10/14/autopilot-automobiles-a-nonsensical-idea/
October 14th, 2008 at 8:21 pm
1) A lot of people seem to be saying that telecommuting is a better option than congestion pricing with regards to reducing traffic, but that seems confused to me, because the whole point of a scheme like congestion pricing is that it would be the thing encouraging more people to telecommute. It’s like if I proposed a carbon tax to reduce greenhouse gases and someone said, “no, let’s just burn less carbon instead”.
2) People tend to talk about robot cars as if the viable technology were 10 years away, but I don’t think that’s the case. From what I gathered as a CogSci undergraduate, the computation needed to allow a car to navigate the road all by itself is the intractable kind, like natural language processing…it’s the sort of technology everyone thought would be just around the corner when computers were invented but has turned out to be totally elusive. But maybe someone who knows more can say that I’m wrong.
October 14th, 2008 at 9:09 pm
In response #13, Eric hits the answer squarely on the head.
Probably _the_ most significant contributor to traffic jams is the propagation of shock waves from drivers braking. This wave’s propagation velocity is related to the speed at which braking is communicated (brake light -> eye -> thinkthinkthink -> foot) through the medium (collective consciousness of drivers on the road), and the entire system of propagating shock waves can be modeled very much like an electromagnetic transmission line terminated into an unmatched impedance. The resultant velocity versus position profile in a traffic jam resembles a standing wave, with velocity peaks and nulls along the highway.
Consider the effect, then, as we reduce the propagation delay to zero (via, for example, a mesh communications network). In this case, the robotic drivers can maintain ideal separation without producing any shock waves (put another way, the standing waves stretch as the inverse of the propagation velocity, so as velocity tends to zero, wavelength tends to infinity), thus helping to eliminate traffic jams altogether.
This argument teases out a significant benefit of robotic drivers: they implicitly, dramatically increase the capacity of existing road infrastructure.
October 14th, 2008 at 10:12 pm
Eric,
The original author (that the linked article is based on) used to be a PRT proponent, and has a whole section on PRT and robocars.
The claim is that robocars have most of the advantages of PRT while not requiring any new construction, right-of-ways or elevated roads. Unlike PRT, robocars don’t require a monopoly, can have different fuels and sizes, can handle sudden changes in demand, and can come right to the houses of people who are mobility impaired.
October 14th, 2008 at 10:35 pm
Congestion pricing would do little to speed the adoption of Telecommuting. The major barriers to telecommuting are not economic – its a cash-flow positive approach today for all stakeholders: 1) it reduces real-estate costs (including parking), 2) it has been shown to increase productivity (less time spent in traffic, more rested alert energetic employees), obviously less money spent on Gas, depreciation…), less traffic, road maintenance costs, etc. The major impediments are cultural – management fears a loss of control, though study after study shows that this is not true (good managers know what their employees are doing, if you can only judge an employees work by whether you can see them, you are in trouble). The outsourcing argument is pretty specious, some of the most successful telecommuting efforts (like Jet-Blue, involve call centers that can easily be moved overseas. the point is that US workers are better at some things.) Telecommuting will clearly be adopted, a concerted push (by government, NGOs, leading corporations, forward thinking Bloggers, etc.) is all that is needed
October 15th, 2008 at 1:23 am
re: david morris
“the computation needed to allow a car to navigate the road all by itself is the intractable kind”
We don’t have to go whole hog to make it useful; in fact, initially robot cars might be dependent on specially-designed roads to guide them, which would easily be feasible in a parking garage, or on a route from the transit station to a parking garage. Robot cars might not be able to mix with human-driven cars for a long time. Agree that entirely autonomous cars are pretty far off, but there are part-way measures.
October 15th, 2008 at 1:28 am
As I say in the original articles which drove this (see http://robocars.net ) well before we can have our robocars, a lot of other technologies will change driving and traffic.
Among those are the fancy navigation systems already being developed for human drivers. These include things like computers that listen to broadcasts of traffic light timing so that you never hit a red light while you drive (because your car tells you it is pointless to go faster now you will just hit the red).
Soon you’ll see broadcasts of both real and predicted traffic for all streets, allowing your system to pick (and reserve) a route for you which load balances the traffic over the entire street network, instead of just where people always go.
Further along, the city could have a traffic control system which redirects streets. In such a system, with 3 streets that are 2-way into town, you could get 2 streets that are one-way into town and one street that remains 2-way, adding 66% more capacity in the rush direction. (Some highways do this now, manually.) Reverse it in the afternoon.
Also, well before robocars we’ll see things like robovalet parking (which we could build now if we wanted) or televalet parking, allowing cars to park more efficiently, and not as near your office as we now demand.
There’s a lot coming from computers in cars even before they take the wheel. Read the original articles for a lot more.
October 15th, 2008 at 6:45 am
Re: Regarding telecommuting: in addition to managers not liking it (they’re afraid their employees won’t really work), employees are cool on it as well because if they can do their work remotely from home, somebody can do the same work remotely from India for cheaper.
Those are issues for full-time telecommuting, but part-time telecommuting (say, one or two days a week) still maintains the idea that being on-site at the office is necessary and gives manager direct access to employees, just not five days a week. I’ve found that this the arrangement that most people who telecommute have this sort of arrangement.
October 15th, 2008 at 7:46 am
The main point about telecommuting is that it can be done today at next to no cost. It can have a significant impact almost immediately and and huge impact over time. Other solutions (robo-cars, bike paths, mass transit) are way-out and require massive investment. Do the math, over half the workforce have jobs where they currently do some work at home (email in the evenings, etc). If you think of half the work force telecommuting 40% of the time (2 days a week), there could be up to a 20% reduction in commuting trips. It ought to be possible to achieve a goal of 25% of that (25% of the workforce working at home 1 day a week). That’s a 5% reduction in commuting, which results in noticeably less traffic and commute times (and gas) for everyone else. (basically for no investment). All of this benefit basically for free. The cost-benefit makes it clear that telecommuting will need to be a major component of any serious global warming initiative. The only problem with part-time telecommuting is there is less immediate direct benefit for the employer, The reductions in real-estate costs don’t really kick in until you go a little higher (3 days a week) and you would need a very strong PR campaign and perhaps some sort of subsidy to get things started.
October 15th, 2008 at 4:11 pm
Hi Matt, I’m the JT who sent you the message for this topic. I wish I’d commented more promptly, since the world moves pretty fast.
Anyhow, the gist of my message wasn’t that I was concerned about increasing traffic (although, yes, it would increase the number of hours that highways are occupied). That’s not my concern at all.
My thought was, if a sizable chunk of the population makes one back-and-forth trip to work each day, and somehow “automated cars” enable the vehicle to head home during the day, that would double the amount of fuel/energy that’s used per car, per workday. Beyond the desire to reduce the emissions and increase the fuel economy of every car, I think it’s simply imperative to reduce the overall usage of vehicles.
It’s appalling enough to see one dude driving along, alone, in his enormous SUV. It’s no less appalling to imagine a car (no matter how small, economical, or efficient) driving along with no one in it.
It’s all a moot point anyway, since driving is an age-old pastime that most people enjoy in one way or another. I don’t see society as a whole ever transitioning to self-piloted cars. We’re too proud
March 1st, 2009 at 11:08 pm
cialis
Excellent site. It was pleasant to me.
March 2nd, 2009 at 9:36 am
levitraExcellent site. It was pleasant to me.
March 11th, 2009 at 9:51 am
Very interesting site. Hope it will always be alive!
March 14th, 2009 at 10:08 am
It is the coolest site,keep so!
xanax
March 22nd, 2009 at 7:42 am
tramadol
It is the coolest site,keep so!
April 8th, 2009 at 4:46 am
I rarely comment on blogs but yours I had to stop and say Great Blog!!
viagra