The link is to a PR piece from a right-wing lobbyist. The actual bill is here.[1] It's mostly about federal preemption. NHTSA can set standards, and states can't. There are also some irrelevant giveaways regarding exemption from bumper and crashworthyness standards for low-volume vehicles.
The preemption part will allow companies to test self-driving heavy trucks in California, something California DMV does not currently allow. Also, currently the California DMV can revoke the vehicle licenses of a self-driving car manufacturer if they do bad stuff, which DMV did to Uber. DMV can probably still do that.
Some of the safety standards are explicitly weak. "The Secretary may not condition deployment or testing of highly automated vehicles on review of safety assessment certifications." But NHTSA still gets to set standards, and they can order recalls.
There's not much about liability; this doesn't change who's responsible for accidents or for vehicle defects. The requirements on manufacturers are mostly toothless - "submit a plan" comes up regularly. There are no privacy standards, so Tesla can watch you in your car as long as they admit somewhere that they do that.
Can DMV still make manufacturers submit crash reports and disconnect reports? Not clear. The bill text is IN GENERAL.—Nothing in this subsection may be construed to prohibit a State or a political subdivision of a State from maintaining, enforcing, prescribing, or continuing in effect any law or regulation regarding registration, licensing, driving education and training, insurance, law enforcement, crash investigations, safety and emissions inspections, congestion management of vehicles on the street within a State or political subdivision of a State, or traffic unless the law or regulation is an unreasonable restriction on the design, construction, or performance of highly automated vehicles, automated driving systems, or components of automated driving systems. Now manufacturers get to litigate "unreasonable restriction". (Some self-driving car companies hate those reports, because they show their technology sucks. Google/Waymo is fine with it. Latest accident report: Uber vehicle just taken out of auto mode was rear-ended while stopped.)
(The biggest lesson we have so far from self-driving car accidents is that the non-self-driving cars need basic automatic braking to prevent low-speed rear-ending the self-driving cars. Google/Waymo cars keep getting rear-ended when they detect they're entering an intersection with blocked lines of sight. They'll advance a bit into the intersection, detect cross traffic, and stop. The human-driven car behind them then sometimes hits them, at very slow speed.)
It's illegal to enter an intersection if unsafe to do so.
The biggest concern having driven around self driving cars in California is that they absolutely cannot handle mixing with traffic in situations like merging onto rush hour freeway traffic. Try watching one get off and watching the gap increase as more and more cars do the same while you are stuck behind the car trying to enter the freeway. It's the eternal good Samaritan problem. Same in normal traffic as it tries to drive 'safely' and maintain a following distance which leads to the same issue of stopping a whole lane.
> Google/Waymo cars keep getting rear-ended when they detect they're entering an intersection with blocked lines of sight. They'll advance a bit into the intersection, detect cross traffic, and stop.
If a human driver did the exact same thing, wouldn't that incite a case of road rage? This seems like a problem that Google/Waymo must solve, rather than forcing the majority of driven cars to solve.
Reducing congestion? Self-driving is going to make congestion much, much worse. Self-driving makes driving cheaper, safer, easier & more accessible. Anytime you do that for anything, usage increases dramatically, often in ways that are difficult to fore see. But it's not hard to predict a few:
- cheap delivery will be used for everything
- people will send and summon vehicles from everywhere. I might drive to work, send my vehicle home so somebody else can use it, summon it at the end of day, and drive home.
- et cetera
While true, self-driving vehicles should be able to optimize the flow of traffic more efficiently than humans do. There are known methods for improving the flow of traffic, such as the zipper merge, that are really hard to get humans to do consistently. As well as the cascading effect of different reaction times to going and stopping, delays caused by too aggressive or too conservative lane changing behavior, etc.
Once you remove the inconsistent skill level, distractedness, irrationality, and emotional reactions that accompany human drivers, you can increase road utilization considerably without a noticeable increase in congestion.
> There are known methods for improving the flow of traffic, such as the zipper merge, that are really hard to get humans to do consistently.
I know a few traffic engineers and one of the things they regularly laugh about are zipper merge advocates (as well as traffic circle advocates, but that's a discussion for another time). It's not that they don't recommend the zipper merge, it's that the advocates think they are magical and ascribe magical effects to them that don't exist in reality.
Want an example of the perfect zipper merge? Look at an hourglass. Thousands of tiny grains of sand perfectly zipper merging at the last instant. And yet there is still a traffic jam. Inflows and outflows still dominate traffic congestion. Now imagine what would happen if you early merged those grains of sand. Your hourglass is now 6 miles tall. Interestingly enough, the early merge is more efficient for flow, but not by enough of a margin (1-2%) to overcome the drawback of the terrible space efficiency.
What the zipper merge is really good at isn't improving flow, it is improving space efficiency. The bottleneck uses far less of the highway when people merge last instant. This allows people to reroute or exit the highway before the traffic jam, reducing the burden on the bottleneck. That's it, nothing more. The idea that traffic would suddenly get so much better if people zipper merged perfectly is not rooted in reality...traffic volume is still the more important factor by several fold.
True, but I think you undersell the benefit of this:
> This allows people to reroute or exit the highway before the traffic jam, reducing the burden on the bottleneck.
Many of those people could have already been intending to exit the highway earlier. Because of the back-up, their journeys are slowed way down unnecessarily. And then you get compounded effects - the highway backs up past another entrance, causing that feeder road to back up past an intersection... In the worst case you end up with gridlock.
I'm not arguing that zipper merges are magical either, but they can definitely have a significant impact.
I don't have any proof, but it's my belief that early merging would work - if people would just merge long before they need too.
Instead, most zoom up until the choke point, then late merge, and rarely do it right, leading to someone slamming their brakes - then traffic jam ensues.
I don't wait for the choke point if I can avoid it. When I see signs that say "construction ahead" or whatnot, I start looking for the signs that say how far ahead, and what lanes. If those don't exist, I try to see what the traffic is doing as far ahead as I can see - looking taillights, turn signals, and anything else - to tell me what lanes are closed, etc.
Then I get over as far and as quick as needed, so I am in the lane I need to be in long before the choke point.
I do the same thing in regular driving; if I know something is going to be on the left side of the street, I try to get over to that side long before I need to - instead of waiting until the last minute and cutting people off (or "begging" with signals to be let in).
I've always found it to be a pain to do it otherwise, even when I don't have the choice and I am forced to, because I have to change speeds and lanes, and that puts me more at risk of collision (it's not the speed that causes problems, but the delta in speeds and lane changes). I'd rather "merge early" (whether that means before real choke point, or just being on the right side of the road to where I am going or turning long before I need to be there) because it seems safer.
The problem is (mainly with choke points, or even on/off ramps) that other people now see the "empty" lane as "I can speed up and go around everyone" and then they get to the choke point and force their way in, or slam on their brakes and swerve, or cause an accident, or anything else. They could see that lane was empty, they had (usually) warning signs miles back saying that the left lanes were closed ahead - but apparently that didn't apply to them.
I don't disagree that it's overhyped, it's just the most recognizable example for me to mention.
> reducing the burden on the bottleneck. That's it, nothing more.
That's precisely it though. By reducing the burden of the bottleneck on the greater traffic network, it localizes the congestion to a smaller area and reduces the disruption the rest of the traffic network has to feel. Overall improving congestion, even if locally the congestion still occurs.
> What the zipper merge is really good at isn't improving flow, it is improving space efficiency.
This is based on the human execution of a zipper merge. Ideally, the continuing lane of traffic will be going at a consistent speed with enough space between vehicles to allow the merging traffic to merge without causing a disruption in the flow of traffic. Theoretically you'd get this benefit from merging earlier as well, but as you mentioned the space efficiency causes cascading impacts to other aspects of the traffic network.
That being said, I only mentioned that as a recognizable example. However, a lot of congestion isn't due to a specific trigger such as a lane ending. It's due to bad human driving habits, and suboptimal human reactions to that behavior. Which create phantom traffic jams[1][2]. Self-driving cars will drive more consistently than human drivers, introducing a level of calm in a sea of inconsistent human reactions. Effectively buffering traffic in a way that will improve overall traffic flow, even before we have majority adoption of self-driving cars.
That's true, but most of those measures require near 100% adoption, which obviously will lag significantly behind introduction of fully autonomous vehicles.
It is possible the reduced need for downtown parking could significantly reduce congestion in cities though.
Will it though? A lot of resistance to more efficient driving patterns is due to emotional reactions (or the fear of emotional reactions). The zipper/delayed merge being the quintessential case[1][2][3]. It works better, a lot of people know it works better, and the only argument against it seem to be "it's dangerous because people are irrational assholes and road rage is a thing"[4]. Once you have self-driving cars around, it might break some of those irrational fears and habits. A self-driving car isn't going to get mad at you if you merge late, it's just going to react to your merging and be done with it. So you're no longer the asshole cutting someone off at the last second, you're just another car that's merging at the point where you have to merge. Road etiquette shifts from emotionally driven to pragmatic, since the self-driving cars around you will also be acting pragmatically. It's not going to slam on the gas and brakes to ride the person in front of them in order to be an asshole to the guy next to them trying to "unfairly get ahead", it's going to maintain a safe driving distance and accelerate and decelerate in a controlled manner to optimize for wear and tear, ride smoothness, fuel economy, and maintaining a safe distance.
Self-driving cars will not only be better drivers on the whole, but also allow human drivers to break through inhibitions related to "proper road etiquette", since it's a whole lot easier to do something subjectively rude to a machine than a human.
That's a pretty big claim with no evidence. Even if people can easily identify self-driving cars and use that to decide to merge, it may just increase their aggressive driving behavior to the point where they cut off self-driving cars at dangerous times. A cutoff that forces the self-driving car to slam on the brakes will propagate backwards and cause more of the fabled phantom traffic jams.
It's conjecture, provided as a counterpoint to the conjecture in the parent post about requiring near 100% adoption of autonomous vehicles in order to see the benefits.
You're completely right that it could cause an increase in aggressive driving behavior. But it's just as possible that it'll decrease aggressive driving behavior. My conjecture's based off of three personal observations:
1) A lot of the aggressive drivers I know that have road rage-like tendencies take it as a personal affront when another driver does something like turn in front of them or using a less congested lane of traffic to get as far ahead as possible before getting over.
2) One of the most effective strategies I had available as a retail manager to diffuse a situation with an irrational/irate customer was to pass blame on to "the system/corporate policy" if they couldn't see reason. They'd still be irate, but they'd redirect their behavior towards something other than the immediate situation. The same principle applies when you go from blaming a specific driver to blaming "those damn autonomous Nissans. They always <insert some peculiar trait of Nissan's self driving implementation>".
3) Aggressive driving is hard to escalate into full-blown road rage when one participant is a computer. As you said, it'll react to the situation by potentially braking aggressively, but it won't then try to tailgate the individual and escalate the situation. It'll back off appropriately and assure that it's not put into that position again and be done with it. Aggressive drivers will get bored with antagonizing autonomous cars when they get used to them and to the fact that you can't get a rise out of them. As well, autonomous cars can likely get better at identifying indicators/patterns of an aggressive driver and account for that.
Lanes dedicated to self driving vehicles could get many of those benefits with relativly low adoption. As adoption increases, more lanes can be dedicated to autonomous mode.
Agree with everything you say. Personally, I hope that self-driving cars are introduced alongside a road-use tax that moves the true cost of road construction and maintenance (and the opportunity cost of all the last currently dedicated to roads) to users, rather than the general public.
Buses and trains are. why should roads be purely subsidized? Why should free parking be subsidized. Car driving is massively subsidized and it has a hugely negative impact on the environment, it's expensive, does not scale, and difficult to maintain. Reducing car use, and shifting to a transit and walking focused style of living is better on every front.
> walking focused style of living is better on every front
Try that in Phoenix, AZ in the middle of July @ 3pm - I'll even buy you a 2L Camelpak filled with ice water.
Trust me, about an hour in you'll be begging for a car.
Snark aside, the problem in the United States (especially in the southwest) is that our cities aren't set up for walking, nor do we have anywhere near adequate public transportation.
That can't be changed overnight (heck, the Phoenix Metro area has only had light rail system for less than a decade); cities won't become magically walkable or bikeable (and I can't imagine Phoenix ever being this way in the summer, unless every employer installs showers for employee use).
There's also the factor that many people (though the demographics seem to be shifting) see vehicles as more than a tool; they seem them as objects of self-expression - almost like a mechanical costume, in a way. There's also the fact that car affords you the ability to get to your destination on your own schedule, not someone else's. Plus, there's just the sheer joy of driving for driving's sake (sometimes, I like to drive my car just to drive it - no particular destination or reason, just because I like to drive - fancy that).
Whether or not car driving is subsidized, it's incorrect to suggest that it's levied uniformly across the population. Gas tax makes up a very large fraction of public spending on roads.
Also, although buses and trains charge per user, they are also heavily subsidized.
Busses are part of the public transportation infrastructure, just like roads, sidewalks, bridges and everything else. It's weird that they aren't fully subsidized.
But roads aren't fully subsidized. That's what the gas tax is for. Drivers of electric cars are currently free-riders, which is something that will have to be addressed if they become a significant fraction of drivers.
Electric cars are technically still taxed on electricity, registration, and purchase. They shouldn't be taxed further until other cars are forced to pay the full cost of their effects including NOx and CO2.
I'm really curious, if it was a la carte, who wouldn't be using it? The OP says it like there's a large group of people not using that infrastructure.. but I can't fathom who those people are?
Hermits living in the woods, alone? Hunting their own food, making their own furniture, etc?
It seems like anyone with even the most minor connection to society are using the goods and services of the road infrastructure. Am I missing something obvious?
If you're making the common argument that they're still "using" it because they patronize stores whose goods arrive by roads and such, that doesn't really make a lot of sense. You want the direct users of something paying for it. If the non-car-owner benefits from the roads through stores using them to ship goods, they should pay for that as part of the price of those goods.
Well, I wasn't exactly making the "argument", but yes that was my thought process.
I disagree though that they shouldn't pay for it. Everyone should. However, the most likely scenario is paying indirectly, in your case.
Eg, if we increased road taxes and applied them only to those who actually use the roads directly, then shipping products will also require price increases to cover said increased road costs. So the end user is paying for the taxes whether they like it or not, in most cases.
Right, but then the end user is paying for it in proportion to how much they benefit. Someone who uses the roads more (directly or indirectly) pays more than someone who uses the roads less. Which is how it ought to be.
The argument that you benefit indirectly could be applied to almost anything. You benefit from the computers used by the businesses you patronize, so should we levy a tax to pay for them? Probably not.
Amount of usage is what's missing from the equation.
Nearly everyone will use the road infrastructure - even indirectly (see: buses, groceries that are delivered by truck), but users (theoretically - back to this part in a bit) have control over the amount of the infrastructure they use.
The ideal argument is that by instituting tolls (or other similar per-use pricing), people will choose to be more efficient in their use of infrastructure, either by shifting modalities (see: taking the bus instead of driving), or shifting travel behavior (see: going to the store near you instead of driving two towns away).
But, of course, this is a highly idealized model that doesn't account for negative impacts of tolling. For example, the poor tend to have disproportionately long commutes, and "consume" more road infrastructure as a result. Tolling negatively impacts this vulnerable group even more heavily than they already are impacted.
Also, the idea that people who can't afford the tolls will switch to travel modes that lessen the burden (buses, trains, carpooling, etc) presumes the existence of competent alternative modalities. This is empirically false in (most of?) America. Proposals for extensive end-user tolling would be less problematic if the revenue can be guaranteed for developing alternate modes. As it stands the vast majority of such proposals I have seen simply institute a cost without offering anything resembling reasonable alternative modes.
To be clear - I'm not 100% opposed to tolling. Around here in NYC there is a plan by MoveNY to institute tolling on all East River bridges which I think is a great idea for raising revenue and reducing congestion.
But the trick is that NYC has multiple competent alternative modes for people to fall back to - the train (both commuter and metro), as well as an extensive bus system - not to mention the ferry system.
But more often than not tolling is proposed without reasonable alternative modes. The poor can take the bus instead of paying the toll - if the bus system was good enough to get them to work on time... If it serves their neighborhood at all.
I'm not sure if the article you linked to is specifically in regards to a specific proposal in LA - I agree with its conclusions, but I don't think you can take a "tolls don't hurt the poor" conclusion from it. Heck, the very bottom of the article literally admits "Tolls may disproportionately burden the poor".
Specifically, the UCLA graphic claims (IMO correctly) "tolling hurts less than a commensurate sales tax increase". This is absolutely correct - sales taxes are regressive and tend to impact low-income people much more than wealthy people. Tolling is the lesser of two evils here, but we can and should do better than just those two options.
There are more ways to fund transport infrastructure than just tolls and sales taxes - property and income taxes, for one, which are progressive and do not have the disproportionate impact on low-income people. Property taxes in particular help the government realize the gains produced by infrastructure investment - a new transit center leads to rising property values, leading to more revenue that can be used to maintain or expand said infrastructure.
In any case, I'm actually for tolling, but only if reasonable alternative modes are already in place. Unfortunately this means I'm against tolling in (most) parts of the US, because frankly alternative modes are either non-existent or so marginal that they are not reasonable fallbacks to driving.
In the spirit of making sure public infrastructure is maximally available and not a la carte, let's make buses and trains free to end users and funded like roads currently are.
But I would like to clarify, I am not 100% against tolls.
Its perfectly reasonable to be expected to pay, for example, a small entrance fee at a national park. But it is unreasonable for this small entrance fee to fund the entire park and unacceptable for it to become a large entrance fee.
Given that everyone relies on vehicles that use roads (emergency vehicles, for instance), there's no such thing as "users" distinguished from "the general public."
The tax would presumably be proportional to (or at least a function of) the amount of use. If the only time a person relied on the road was when they needed an emergency service, then they would pay very little tax indeed. If someone decided to start a delivery business, which saw thousands of cars going to and from peoples houses every day, then they would pay a much larger tax. This tax would presumably be passed onto their customers, making their customers pay a higher tax than the rest of the public.
Shouldn't "users" include everyone who benefits from the roads being used to bring things to them, like groceries to the local store and laptops from manufacturers?
I assume you are in favour of financing public transport also from the pockets of only those who use it? Maybe libraries, too? How about schools?
(Truth be told, I'm from quite the opposite land where driving taxes finance much of the public spending. I wouldn't really mind your proposition too much.)
Disagree entirely. Yes, the number of cars on the road and trips driven will increase, but this will not overshaddow the traffic advantage from SDCs.
All in all SDCs should increase road bandwidth about 5x by my estimates.
SDCs can drive much, much closer together. They will also never be the cause of phantom traffic jams. For every human that switches to SDC, phantom traffic jams get less likely - a main cause of traffic - which again increases road capacity.
We don't have to wait until 100% SDCs to start increasing road capacity either. SDC "convoys" travel as a single unit, and interact cleanly with human drivers.
Regardless of SDCs, when road capacity is reached, traffic will happen. But SDCs have the potential to greatly increase road capacity.
No they can't, not if you want them to be safe. The biggest problem we have with human drivers, most people will say, is that humans can't drive safely. And one of the biggest reasons that is is because we drive way too close to each other. In an urban environment, it is practically impossible to maintain safe following distance: someone will slip in and take that spot instead.
Sure you might say that they can communicate with each other--they can't, not in any reasonably close timeframe for infrastructure (say, the next 30 years) where cars have to assume that human drivers are still on the road.
A worse effect is that self-driving cars are more likely to strictly stick to rules of the road and safe driving than humans are: so longer following distances, slower speeds, strict right-of-way rules when trying to make a left onto a jammed road... I don't see how they're going to be better.
But about half of the necessary safety margin comes from reaction times. Computers tend to be much better than humans, so that autocars can get closer without compromising safety.
> With only self driving cars on the road we have the ability to fix #2, and thus greatly increase road capacity and reduce traffic. Parity would be reached at, say, 4x more cars.
This is fantasy land. This idea has been perpetuated by a few layman websites and extremely simplistic closed-loop simulations, but have never been validated by the more sophisticated traffic simulation systems used by traffic engineers. In fact, quite the opposite.
Bad reaction times contribute to waves, but waves are just annoying and feel like the cause of congestion to the driver. According to the traffic engineers I know:
1) You can't ever fully get rid of traffic waves because reaction time is only a fraction of the speed mismatch correction problem, physics still dominates.
2) You can't ever fully get rid of speed mismatches. Even a plastic bag flying on the roadway is enough to cause them.
3) Traffic waves only feel like congestion to those who experience them first hand. Minimizing traffic waves doesn't decrease travel times significantly, it merely moves congestion to harder bottlenecks like traffic signals, highway on and off ramps, or merge points.
4) Braess' paradox only manifests itself when agents are acting rationally. As rationality increases, whether through better mapping and routing systems like Waze or by algorithmic rationality of SDCs, more bottlenecks appear that didn't previously manifest themselves under a lower rationality state. Unless you can somehow centrally control SDCs (which is unlikely IMO), this will only become more prevalent.
5x seems extreme, but I'll certainly grant you 2x on the freeways.
But if you put twice, or worse 5x, as many cars on the freeways, that'll just jam up the urban core even worse than it already is. And self-driving cars are probably less efficient than human-driven cars in the urban core simply because they're much less aggressive in their interactions with pedestrians and cyclists. This is a good thing, but it will increase congestion.
Efficiency improvements can only add increase capacity by a fixed percentage. The massive improvements in cost ($0 in labour!) and convenience will cause usage to increase by orders of magnitude, IMO.
They will be driven more efficiently. People are VERY inefecint with driving.
Self driving cars won't:
1. Increase their space when the traffic slows down to where there is 300 feet between cars.
2. Less accidents to cause shut downs and lane closures.
3. Merge properly instead of slamming on the brakes and putting on their turn signal 1/4 mile from the merge point. Merging will happen while moving instead of from a stop.
4. Touching their brakes even though they aren't slowing down.
I'm not sure if those would lead to really major changes to congestion, specially once you consider that there will be increased demand for driving once self-driving cars are ubiquitous. The only approach to truly reducing congestion is to switch towards denser public transit.
BTW, I suspect that the bigger impact of self-driving cars would actually be the reduced need for parking space.
Quite funny you'd go as far as capitalize your emphasizing, considering that people are the only group of drivers able to drive everywhere.
Touching your brakes to flash your brake lights when braking with your engine isn't an altogether bad idea on congested road with traffic waves – it's a good signal to the car behind that you're slowing down (before they can actually see that) so that you can together help mitigate the propagation of the wave.
That is a valid reason to flash your brake lights. I do that, particularly when I know I'm slowing or will be slowing and someone is following too closely or can't see what I see (that there are brake lights being hit 5 cars ahead and my car is too big for them to see around/through). But I do my best to be a courteous driver.
Here in GA, most people don't even seem to understand that the left lane on the highway is the fast lane. Every time I'm on I-75 for more than 5 miles I'll be going speed limit to speed limit +5 and get stuck behind someone doing speed limit - 10 in the left lane. I do not believe that the majority of people tapping their brakes are doing it for a valid reason. It's more often (based on them accelerating at the same time) because they're riding their brakes.
Mathematics has shown that each time a brake light flashes it forms a wave that miles behind causes total stoppage because for 2 minutes people hit their brakes lightly slowing down with each person behind slowing a little bit more.
I touch the brakes only when I'm slowing down, but not necessarily to cause me to slow down. And I only do that when people are following too closely behind me.
I've been rear-ended by people following too closely who failed to recognize that I was slowing in time to slow themselves down. I'd rather be pragmatic and warn them off before an accident occurs.
By "will be slowing" I mean: I have removed my foot from the throttle or stopped cruise control but my vehicle has not started to appreciable slow yet.
Yet on the other hand many people have proposed parking will be eliminated- instead of parking your car, you'll just send it off to drive in circles around the interstate for a few hours! Wouldn't that be marvelous!
Maybe it won't be your car, but part of a fleet. So it drops you off and then picks someone else up, perhaps someone doing the reverse commute. Even if it has to go park itself at some fleet lot outside the city center, it should be a marginal improvement over circling, looking for a spot.
You'll change your mind about it not being your car the first time you get into one, it takes off, then you're like "sniff...sniff...ohgodwhatzatsmell", suddenly realizing some drunk person puked (or worse) in the back seat area.
That's one reason I avoid public transit in San Francisco.
However, it hasn't been a problem in my use of taxis, rental cars, Uber, or Lyft, so I presume it won't be a problem with autonomous cars. They'll either drive themselves to the car wash when someone has an accident (spills, pees, vomits), or if for some reason a "wet" car arrives, then you'll report it, and get the next one.
And park closer together. I imagine my office parking lot's physical capacity would increase by at least 60% if we didn't have to get in and out of the vehicles. It already has capacity for around 1000 workers (maybe more).
Even if they didn't move the lot, my car could drop me off at the entrance and then go find an available spot. The lot is huge, close to a 1/4 mile walk from the furthest available space to my entrance. I wouldn't mind not having to worry about that on those days I'm running late.
And they won't need as much space for backing/driving- especially if they can communicate with each other- a car that is blocked in could signal the other cars that it needs out, and they could move to make space for it to get out.
But those cars can drive in tandem. When humans drive, every start and stop has a delay between each driver, and inattentions compound those little delays. Additionally, traffic caused by rubbernecking, by trying to cheat lanes, by not zippering, by reaction inequities all are human problems.
This is what Americans want. Anti-car zealots need to stop thinking you're going to pack the entirety of the country into trains and buses. It's not going to happen, so stop trying to make things painful for the 80%+ of Americans who drive, want to drive, and are interested in government making it easier and more convenient for them to drive.
I cannot possibly fathom how the left (which I usually include myself!) went from loving the idea of self-driving electric cars to "no! no! we need to stop people from driving at all costs!".
Cars are also unused for over %90 of the time, taking up space in some parking spot on premium land close to people's workspaces and homes. On almost every road in america, 2 lanes of traffic are taken up by car parking. In a self driving car future, we can free up those lanes and put excess cars in parking / maintenance towers. The software system managing can predictively dispatch them based on previous demand curves to prevent traffic jams.
The amount of cars we would need as a society will equal the peak demand curve, much like electricity demand today. It will also be very likely that those cars will have multiple people in them since a large part of the cost savings will come from car sharing. In the future, we will probably have 1 motor vehicle serving 10 households, vs the almost 1:1 to 1:0.5 ratio we have here in the US today.
Exactly, I don't understand why people think self driving cars will do anything to reduce congestion -- if anything they are going to make it worse since they make point-to-point car trips easier and cheaper, and instead of driving to work and parking all day, the cars will drive someone to work, then drive somewhere else to get another rider (or park to await the evening commute).
They may bring some efficiency to traffic patterns leading to better road utilization, but that only goes so far to make up for the much higher level of traffic that will result when it's much cheaper and easier to call a car when you want one.
San Francisco is already feeling the "Uber effect" of greater traffic from car-share services:
Everybody loves the idea of new technology that allows them to continue doing what they are doing, in this case avoiding biking, walking, and public transit. Basically, people want to believe self driving cars will reduce congestion. Ergo, it's pretty easy to convince them it will.
That "Uber effect" is a circular citation. SFGate is reporting that SFMTA says Uber/Lyft are causing traffic (SFMTA has an agenda), and in their 'report' they cite SF Chronicle (the 'newsier' site of SFGate) as evidence for the claim that Uber and Lyft are to blame:
That article provides no evidence to support this claim other than the fact that 37k people have driven for Uber/Lyft/et al in SF. The author then extrapolates that adding 37k drivers to SF is a huge number. As if all those drivers appeared in SF ex nihilo.
Is Uber really comparable? With Uber each vehicle is trying to be in the busiest place to get a fare operating only in its own self-interest. With an autonomous fleet they will be where they need to be and will get out of the way on their own.
Why would the motivations of a driverless car company be more altruistic than that of an Uber driver? An idle car makes no money, fleet owners will send their cars where they are more likely to get a fare.
Even if the cars are privately owned and people just have their car come in to drop them off at work before sending them out of the way to park somewhere (since inner city parking is scarce, so there's no space to park every car at the worker's building), that doubles the number of trips -- once to drop them off, and once to go somewhere and park.
Fleet owners will send as many cars as they need into the fullest places, and as many cars as need into the emptier ones. Uber drivers are each one out for themselves, so they can't do any global optimization.
But a company will not put two cars in the same place to compete for one fare.
And the remote parking may double the number of trips, but the parking is not going to be far away and will take much less space than a normal parking lot.
In most of America, this idea seems to be even less practical than self-driving cars. America is pretty much designed to be a pessimal case for public transit. San Diego, where I live, does invest in public transit infrastructure and has relatively extensive mass transit options (buses, trains and trolleys), and yet public transit is still largely unusable outside of fairly small, specific areas. In order to use public transit, I'd have to walk half a mile to a station and then my commute to work would be four times as long as it is by car.
So when I hear "invest in public transit infrastructure," I feel like it isn't a concrete enough suggestion. Does it actually mean tear down most of America and replace it with a country where mass transit is practical, or what? I'd love never to have to drive again, but merely investing in public transit demonstrably isn't enough.
The issue with San Diego is not the lack of infrastructure, but the land use— many businesses and offices are in unwalkable office park hellscapes that are designed to make driving easy to the exclusion of any other mode of transit.
Since public transit requires a walking trip at both ends, it will never be very popular in areas where one needs to take a dangerous mile-long walk across a sea of parking and high-speed arterials, instead of a quick jaunt through a lively neighborhood.
That was exactly my point: A lot of the US is laid out in a way that is pessimal for mass transit — "hellscapes," as you put it — so investing in public transit infrastructure doesn't seem like enough to make a big difference. The problem isn't just a lack of investment, but also that mass transit isn't very well suited to America as it exists today. Any proposal to make public transit happen in America that doesn't address this seems pretty unrealistic to me.
San drego has a surprisingly good train system with the light rail and coaster. It's just sitting on one of the most ass-backwardsedly designed cities out there. I can take the trolley to Santee, but not the airport?? not south park or north park (the whistle stop on thirtieth is named for the steetcar stop it used to be!!!!)? not Balboa park? It's a disaster of a city and the trolley deserves better.
America is low density because it provides incentives for low density living. The solution, in short, is to remove those incentives. Once the costs of current land use patterns are undistorted, society will adjust, and land use will intensify.
I guess investment in public transit has to go hand in hand with saner zoning laws and measures to reduce sprawl. It's certainly not an over-night solution, but few things in city planning are.
Agreed! I would love to look up directions to a destination, and find that the driving and transit time are comparable. As it is now, it's laughable: I live in SF and my SO lives in Oakland; driving time is usually 18-25 minutes, but transit is 50-70 minutes, depending on when I leave.
Here there are two problems:
1) "Last mile": it's a 20 minute walk for me to get to the right transit line, and then 10 minute walk after I get off BART. The BART ride itself is only about 20 minutes; my "transit" use is mostly not actually on transit.
2) Timing: during most of the times of day when I'd make this trip, BART runs around every 20 minutes. If I'm one minute off, then I've spent 50 minutes walking or standing around for 20 minutes of actually being on transit.
We need more transit lines, going to more places, and more-frequent trains during the entire day.
It seems like there's always a "we're going to try it here" phase, then it never goes further.
The upsides sound great - basically personalized mass-transit, using individual pods on an overhead track system. The pods and track are lightweight, and can likely be put into tighter routes and areas than a full-sized train can go, as well as not disrupting traffic or whatnot when it is installed.
It does have the whole "monorail" stigma to get around (which I don't really understand why or how "monorails" got the stigma - they have issues, certainly - but the whole meme seems to only have started since that Simpsons episode)...
I would consider that as long as public transportation evolves. Light rail can make a lot of sense, but I have a hard time getting behind buses. I rode one for years to get to work and I hope I never have to again.
If instead public transportation worked more like Uber or Lyft I could get behind that. I wonder if some cities with underused buses wouldn't be better off subsidizing their citizens use of Uber?
Carpooling is made much more viable by self-driving cars. Could easily see people being willing to share a ride if the computer matched up people from the same neighborhood for a ride automatically and the cost was shared.
Waze Rider and Scoop are actually pretty great for matching people up for carpools (and are specifically for that purpose and that purpose only, unlike Uber Pool/Lyft Line), and yet I don't see a large number of people using either service.
I suspect people with cars are still very protective of their ability to completely control their schedule; agreeing to ride in someone else's car makes it difficult to run errands after work without first stopping at home to pick up the car, for example. And what if they have a ride scheduled for 5pm but find out at 3pm that they need to stay later? With enough scale it shouldn't be hard to find an alternate ride, but we're not to the point where people can expect that to be easy.
I think people could definitely rearrange their lives a bit to make this better, like planning to run errands on weekends, or close to home (instead of work). Or setting reasonable boundaries with their employer (I'm scheduled to leave at 5pm, and not a minute after -- unless you want to pick up the tab for a taxi). But people are selfish, and stuck in their ways, and would, for some reason, prefer to sit in traffic.
There's this innovative type of carpooling called a 'bus', where you can fit more than 30 people onboard at once! The cost per person is incredibly low, too!
At this particular moment, according to Google Maps, it would take 2hours 5min to get to SFO by a mix of bus and train, or 23min by car. Bus/Trains are cheap, but when you have more than 30 people on board you probably have to make a lot of stops that slow down the trip.
Not really. In reality, even the buses of England for example end up driving so many trips relatively or completely empty (to provide adequate service so that you could do without a car) that they end up producing more emissions per kilometre compared to personal vehicles – you can't do that with incredibly low costs, nor can you calculate the costs by thinking of only the full buses, when that's hardly the complete picture.
It's unclear how pouring even more money into incompetently managed, failing public transit systems is going to actually reduce traffic congestion. For example, let's say we increased BART's budget by 10X. How exactly would they use it to "reduce congestion"? Are they going to build 10X more routes? 10X more stops along existing routes? 10X more trains serving existing stops? Or are they just going to funnel the money into 10X more bureaucrats' pockets?
While your cynicism isn't entirely unwarranted, I think there's more to "investing" in public transit than just throwing money at the problem. Ensuring that the money actually goes to good use is part of it. Not sure how we do that, though.
I would say even the minimal change of funding 2x as frequent trains on BART would do wonders for making it more useful. Unfortunately residents in the bay area have way too much say in what does and doesn't get built, so new stops and new lines are difficult, to say the least.
Don't get me wrong, I'd love it if there were actually a way to make public transit better! If I could send my money somewhere and know that as a result, I could go places by bus faster than I can by driving myself, my check would be in the mail tomorrow. If I could vote for some politician and as a result, BART would come out to my city (Livermore), I would do so. But nobody is promising any of that.
If you convince enough people to invest in self-driving car companies, eventually self-driving cars happen. Voters just approved a $3.5 billion bond measure for BART. They're not even using it to expand service--just to "maintain existing infrastructure". They need billions just to stand still.
Sure, I get that and agree. I just object to the idea that "investing" in public transit just means "throw money at the transit agency with no plan and no accountability".
> Voters just approved a $3.5 billion bond measure for BART. They're not even using it to expand service--just to "maintain existing infrastructure". They need billions just to stand still.
Yes, and that's what you get when you underfund something for decades. Technical and maintenance debt piles up, and it costs increasingly greater amounts of money just to hold the status quo. I don't know for a fact that this $3.5B couldn't be managed in a better way than planned, to not only maintain what we've got, but also improve, but it doesn't seem that far-fetched to me, either.
I'm not saying it's wise, given how BART is currently managed, to throw (say) $10B at them, but, managed properly, perhaps it'd be enough to not only get them out of the hole, but also make actual real improvements? The trick is, of course, giving them that money in a way where we'd actually see good results, and I agree that, without other things changing, we probably wouldn't get much for that money. At least not enough to be worth it.
In the end it's the way the bay area is structured politically that's the problem. BART can't build new lines because the citizens along the way are all of the NIMBY mindset, and our governing structure gives them too much power, and politicians are too cowardly to override them. The Fremont->Diridon expansion plan has been on public maps since before I moved here in 2004; and we just finally got the Warm Springs expansion earlier this year. And who knows about San Jose. The delay isn't because BART sucks at building or managing construction projects, it's because the people and municipalities along the track path are selfish.
> I'm not saying it's wise, given how BART is currently managed, to throw (say) $10B at them, but, managed properly, perhaps it'd be enough to not only get them out of the hole, but also make actual real improvements?
For comparison, Elon Musk built a space program, orbited the Earth and docked a rocket with the ISS for 1/10 of that amount.
I agree, it's not lack of money that's holding public transit back.
Certainly building a reusable rocket in a mostly greenfield environment isn't quite apples-to-apples with maintaining a decades-old transit system that is politically difficult to maintain and expand, but it does illustrate the point that amazing things can be done on a small budget.
Privatizing public services is often fraught with peril, but the bay area transit systems have proven that they're incapable of meeting needs with current management. Not sure what we can do here, unfortunately.
The title could do a lot more to indicate what the Self Drive act actually is.
A house committee drafted some legislation that grants the federal government regulatory power over self-driving cars, or rather strips it from state governments. At least as far as it is reported in the article.
It'd be interesting to see what the reaction is from people and companies in the field.
Preemption legislation seems a little premature, to me. Sure, it would stop states from hindering autonomous-vehicle development, e.g. to protect entrenched interests or because of irrational fears, but it could also stop states interested in being on the forefront of research from doing more interesting things.
It seems like the sort of thing you'd want when products were further along the development path and having a broad, harmonized-rules marketplace is necessary to move further, but not when most of the tech is still pretty early R&D work.
Preemption is needed most for nascent industries where the economics are uncertain. It gives investors more confidence to invest in an unproven technology if you can at least predict one aspect of the landscape on a national level.
Down the road, you can start to relax the preemption (theoretically) and let the market forces keep unwise state regulation in check.
I skimmed the text, it does specify requirements of it's own on self-driving cars (both for safety and privacy), establishes a council of industry experts on it, etc. as well.
But barring states from regulating this in their own borders is about as much of a sweetheart bill for business as you can get, it ensures they only have to spend their lobbying money on one legislative body.
But barring states from regulating this in their own borders is about as much of a sweetheart bill for business as you can get, it ensures they only have to spend their lobbying money on one legislative body.
But it also means that they only need to meet one standard, not 50. Which sounds like the right thing to do when it comes to regulating something mobile that will operate in all 50 states.
The differences from state to state are typically minor and not incompatible. Most states derive their requirements from federal standards/recommendations anyway.
Current manufactures only offer two versions of cars. One is for most states. The other is for a particular state and all the other states on the "I'm with him" bandwagon.
If a certain state would adopt federal recommendations for certain vehicle requirements then things would be greatly simplified.
Even minor changes between states are expensive for automakers to manage, and it's bad for consumers if some manufacturers choose to certify their cars for certain states -- you could end up buying a car that you not only can't legally take to another state, but the car will refuse to go there at all to avoid manufacturer liability.
There's very few situations where a car wouldn't be legal to drive in another state. For one, states are generally expected to respect other state's certification and licensing. Additionally, short of the hallmark "cars which have no steering wheels", which likely just won't work outside their programmed area anyways, you could just drive the car normally when in a state which doesn't support it.
You want to reduce congestion and make mobility improvements for many? Look to cities that make cycling a way of life through infrastructure, planning, and design.
I really hope that's just wishful thinking from companies like Qualcomm and wireless operators and not something car makers are actually considering.
Making the cars' critical systems (such as the self-driving systems, which would respond based on other cars' actions in car to car communication) be accessible from the internet sounds like a terrible idea. This is why I hope this type of law is not rushed, as car makers and companies like Uber and Google/Waymo hope it will be.
There's a lot of potential gain from car-to-car communication, but I also see a huge risk. I can be reasonably confident in my car's own sensors, but only God knows what my car is being told by other cars about the world; the opportunities for malicious activity there are huge.
It's hard to imagine a system that uses V2V that doesn't rely on some manner of trust, like a closed/encrypted system with car manufacturers having CA-like signing authority or something. And if I've ever learning anything, it's that car companies are not great at network security.
I'm confident in many benefits of autonomous vehicles, but reducing congestion is not one of them. Quite the opposite, actually.
Wanna know what happens when you no longer have people looking for parking but rather have them picking up and dropping off at the curb? Visit a busy airport or pickup/dropoff curbs at a large suburban elementary school.
Now throw in the fact that in order to get to your destination, you'll have to drive through those clusterfucks that are caused by other people's destinations. Oh, and all the cars that might not be circling looking for parking, but are now circling looking for new passengers. Oh, and all the new cars on the road when lower costs move people from transit into SOVs. That's the future, embrace it.
Sure, if we accept the idea that traffic congestion increases, but I don't think that it will be the biggest problem because with current forms of congestion it becomes frustrating because you have to pay attention to the road and operate the car, while alternatively with a SDC, that'll no longer be the case and can actively engage in productive activities (or not).
I suppose there are other arguments to this; Maybe time critical events such as emergency vehicles attempting to reach a destination, but there's probably a solution to this...
It's like introducing a new more efficient protocol over existing limited networks. It can reduce congestion sure by using the network more efficiently, but only if done cold turkey (no or limited human drivers). This is unlikely to happen in ge states, but much more likely to happen in Asia (China) where they have real traffic problems with hard limits on infrastructure.
> but reducing congestion is not one of them. Quite the opposite, actually.
The premise of that is to increase vehicle usage efficiency, hence reduce the total amount of vehicles on road and reduce congestion.
Additional autonomous vehicles applied in public transits will further improve traffic flow (on demand vs. fixed scheduled services, reducing empty runs etc.)
Some drivers need a space fit for a semi truck in order to pull out into traffic.
Some drivers don't go at green lights because they're distracted.
Some drivers camp in the left lane.
It's highly possible that the cumulative benefit of widespread adoption of self driving cars could offset the negative aspects of people who suck at some subset of driving.
Perhaps cars will be smaller? If people choose to take the cheapest available self-driving taxi, which happens to be the smallest car that will fit their need, cars may end up smaller and less space will be taken up by vehicles on the roads and in such loading zones.
It's an interesting problem for the prestige car manufacturers, people don't care what the make of a taxi is in the same way they do about the car parked on their drive.
There will be more efficient algorithms developed over time, all it will take is a software update - behavior will change overnight (for better or worse).
Will there be a lot fewer cars, though? Assuming you want to achieve the same throughput, you'd still need about as many cars as there are at rush hour today.
25k is less than 1% of most major auto manufacturers yearly production numbers. No article I've read on this news has put that into context and instead praises the legislation as liberating car manufacturers to bring FSD cars to the masses. Most people don't read past the headline let alone look into the facts of the article. Also nothing good EVER gets passed legislation unanimously. Last time the House passed legislation this quickly and unanimously was when they passed SOAPA. Something feels fishy about this. I haven't read the legislation directly but I bet if a journalist went through it fully, somethings would surface that people wouldn't like.
The House Energy and Commerce Committee passed the SELF DRIVE Act with a vote of 54-0.
The act correctly delineates the purview of federal versus state regulation for autonomous vehicles. In short, federal regulatory bodies have authority when it comes to the car, while states have authority when it comes to the driver.
A bipartisan, unanimous vote to secure future regulator power, story at 11! eyeroll The Federal government gets to regulate the car that drives itself, and the State gets to regulate the driver that's actually just a passenger.
I don't get this assumption that SDCs are going to reduce energy usage. Sure, it will be better on a per-mile basis, but people are going to be traveling a lot more than they are now.
It may not, but hopefully it is an opportunity to push the transition to electric cars. If we are going to effectively replace all cars, let's make some smart choices on the replacements.
A new designated felon would be required to sell highly automated vehicles, the Cyber Security Officer.
You'll sign some documents certifying everything is fine, but of course, everything won't be fine. When your cars get hacked you'll go to jail for defrauding the government and/or manslaughter. I bet it pays real good though.
Does it give them permission to remove the backup driver? In Waymo's case it seems like they are mainly waiting for a law that says that and then they can deploy for certain routes and conditions.
The idea that a bill "sponsored" by corporations whose sole purpose is selling as many cars as possible will "reduce congestion" is probably the most ridiculous idea this turn of the century.
What's next? McDonalds innovation on nutrition facts labels to reduce junk fat consumption and improve health outcomes?
If corporations selling cars were "sponsoring" a bill, they should be sponsoring a bill banning self driving cars, because they are an existential threat to the business of selling cars.
The preemption part will allow companies to test self-driving heavy trucks in California, something California DMV does not currently allow. Also, currently the California DMV can revoke the vehicle licenses of a self-driving car manufacturer if they do bad stuff, which DMV did to Uber. DMV can probably still do that.
Some of the safety standards are explicitly weak. "The Secretary may not condition deployment or testing of highly automated vehicles on review of safety assessment certifications." But NHTSA still gets to set standards, and they can order recalls.
There's not much about liability; this doesn't change who's responsible for accidents or for vehicle defects. The requirements on manufacturers are mostly toothless - "submit a plan" comes up regularly. There are no privacy standards, so Tesla can watch you in your car as long as they admit somewhere that they do that.
Can DMV still make manufacturers submit crash reports and disconnect reports? Not clear. The bill text is IN GENERAL.—Nothing in this subsection may be construed to prohibit a State or a political subdivision of a State from maintaining, enforcing, prescribing, or continuing in effect any law or regulation regarding registration, licensing, driving education and training, insurance, law enforcement, crash investigations, safety and emissions inspections, congestion management of vehicles on the street within a State or political subdivision of a State, or traffic unless the law or regulation is an unreasonable restriction on the design, construction, or performance of highly automated vehicles, automated driving systems, or components of automated driving systems. Now manufacturers get to litigate "unreasonable restriction". (Some self-driving car companies hate those reports, because they show their technology sucks. Google/Waymo is fine with it. Latest accident report: Uber vehicle just taken out of auto mode was rear-ended while stopped.)
(The biggest lesson we have so far from self-driving car accidents is that the non-self-driving cars need basic automatic braking to prevent low-speed rear-ending the self-driving cars. Google/Waymo cars keep getting rear-ended when they detect they're entering an intersection with blocked lines of sight. They'll advance a bit into the intersection, detect cross traffic, and stop. The human-driven car behind them then sometimes hits them, at very slow speed.)
[1] https://www.congress.gov/bill/115th-congress/house-bill/3388...