Trump administration reportedly kills vehicle-to-vehicle safety mandate

Is this finally the end of the road for V2V?

Read the whole story

 

[Fatesrider]

Good. V2V is troubling for privacy as it makes automated tracking even easier than plate scanning. It also won't work with motorcycles that tend to avoid these fancy features, bicycles, pedestrians, and other obstacles. As general obstacle avoidance technology is basically ready for production, is there even a need for V2V?

here's the thing if vehicles have cellular modems in them then insurance companies can monitor everything you do with the vehicle and if you do one wrong thing ever then adios insurance.

Only if the insurance companies have access to that data. I don't see any requirement to give it to them on the horizon.

What do you mean "requirement to give it to them". It's fucking wireless. If they want it they can get it whether or not YOU want them to, whether or not there's a requirement and whether or not you even KNOW about it.

I mean, really?

And for the record, the same thing can happen any time someone plugs into your car (say for maintenance, diagnostics, smog checking, etc) and all that can be uploaded to a DMV database the insurance companies would have easy access to, all nicely sorted by VIN.

So, it's not like you're immune to it even IF there's no wireless.

Isn't that a pleasant thought? Welcome to the 21st century where your privacy is company business.

 

[Criscokkat]

And as I pointed out, when self-driving cars are the norm, you won't have a choice. Not many human drivers will be able to afford the insurance after risk assessments of human drivers vs AI drivers are available.

Even if you don't have a cell phone, you are being tracked with license plate readers and CCTV. You're not going to roll that back.

You can't stop this technology revolution by not carrying your phone with you when you drive. It's inevitable. The only thing to be determined is whether V2V saves lives. People were complaining about having to wear seat belts as an invasion of privacy and that argument didn't work then and it won't work now. When self-driving cars are the norm, you're going to lose every last vestige of privacy while driving anyway.

Will motorcycles be self-driving too? I don't think that would be practical. Will they be banned?

I don't think motorcycles will be self-driving for most - kinda kills why people buy them. However this doesn't mean that they couldn't be part of a v2v network too, especially for informing other cars where they are and what their speed is. If in the future the auto companies come up with a workable consortium plan on how to communicate this information I could see cell phones that broadcast this information to cars around them when they detect their users are walking or biking as well.

 

[Elsmarc]

Good. V2V is troubling for privacy as it makes automated tracking even easier than plate scanning. It also won't work with motorcycles that tend to avoid these fancy features, bicycles, pedestrians, and other obstacles. As general obstacle avoidance technology is basically ready for production, is there even a need for V2V?

here's the thing if vehicles have cellular modems in them then insurance companies can monitor everything you do with the vehicle and if you do one wrong thing ever then adios insurance.

Only if the insurance companies have access to that data. I don't see any requirement to give it to them on the horizon.

Anyone who really thinks this isn't coming in the not so distant future one way or another is deluding themselves . As a youth I could not have conceived of TV cameras mounted throughout cities, much less facial recognition, or "smart phones" for that matter. Not to mention the NSA, the internet and oh, so much more.

There are already TVs which technically can "spy" on people in their homes. And how does one really know if their "smart phone" is spying on, watching and listening to them?

With the unbelievable (my opinion) acceptance of Siri, Alexa and similar voice (and soon to be video) "assistants", I can only see acceptance of a future where it will be quite normal despite a small vocal minority of opposition.

Just look at credit "bureaus". About a year ago I sent a request for all "my" information to the "Big 3" credit companies, as well as LexisNexis and 2 other companies. I got several hundreds of pages of "information" and was shocked, to some degree, at how much "they" know about me going back many years. Then gain, most of it wasn't surprising (and some of it was totally incorrect).

And I remember back around 1990 when I was buying a house an insurance company declined to cover me. It turned out they had done a credit report and gotten some "bad" (incorrect) information. Once that was sorted out I got a very good rate.

Very few people know how much information companies have on them today, and it won't change to less information in the future. Get used to the future. Oh - And don't ever go into a casino - Your face and quite a bit more will be distributed to casinos, police organizations, various "crime prevention" companies and a lot more.

 

[Bluestrike2]

"Hopes have long been pinned on V2V as a way to cut traffic fatalities, which have been on the rise the past two years. "

Traffic fatalities have been steadily rising ever since the introduction of texting on a smartphone which are now used by morons simultaneously while driving. THAT is one problem that really needs addressed ASAP.

Unfortunately, distracted driving doesn't really have a decent technical answer.

It's not like car companies can turn cars into faraday cages. Even if they developed shielded EMI automotive glass, drivers could just roll down a window. Phone manufacturers can detect that the phone is moving, and that it's likely in a vehicle. They can't easily detect whether you're a driver or a passenger, or if you're on a bus instead. If auto manufacturers started including iBeacons in their cars, I suppose you could maybe create an internal positioning system in the car to "guess." But that'd involve the car industry standardizing it (layout, tech, etc.) and writing a protocol to transfer environmental context information to the client [re: phones in the car] so they could use trilateration to position themselves within the car. Even then there would be a lot of guess work. Many people mount their phones on their dash, often in the center of the car or even slightly to the right. That's technically part of the "passenger seat area," but should we classify it as "likely the driver?" You'd have relative probabilities, rather than definitive answers. And that introduces all sorts of liability for the phone manufacturer when it comes to drawing a brightline to determine whether a phone should be locked down.

Of course, there's a nifty hack available: interrogate the occupant classification system (OCS) sensors to determine if there's a passenger in the front seat. If there isn't, you know the user is the driver and is just itching to use that phone. Lock it down! Of course, if there *is* a passenger, you're right back where you started. And then there's the back seats, which don't have OCS sensors since they don't have front airbags to worry about. You'd also need to build the hardware and write a standard for broadcasting the OCS determination in the car as well. Darn.

The simplest solution would be to detect if the car is moving, and then force a prompt asking if you're a driver and a warning in nice big letters. It's also the least helpful solution. People lie. They even lie to themselves. Besides, we already have prompts like that and big, ugly warnings for dedicated navigation systems that are ignored (though in-car systems can just lock you out entirely when not stopped).

Most states have distracted driving laws on the books: 15 ban hand-held cell phone use altogether; 47 plus D.C., Puerto Rico, Guam, and the U.S. Virgin Islands ban text messaging; all but 2 states note distraction on crash reports when they're filed. Relying on people to follow those laws, or honestly answer a prompt asking "are you a naughty driver" means relying on the honor system. In this context, that doesn't work.

It doesn't help that there are some valid scenarios where people need to access their phone in the car. If your solution locks down a passenger's phone and accidentally prevents them from calling 911, there's the potential for great harm in that individual situation. Yes, it's potentially outweighed by having fewer distracted driving incidents elsewhere. But that's cold comfort to the people in that scenario. And regardless of what technical solutions are developed, there are still 253 million other cars on the road with an average age of 11.4 years that won't have them.

Integrations like Android Auto and Apple Carplay are basically an admission that you can't stop users from wanting to use their phones. If you can't stop them, you might as well give them a means of interaction that minimizes how much distraction it causes. That's accomplished through simple UI design, easy-to-use voice control with natural language, reading back messages, etc.

Distracted driving is absolutely a problem. The distracted drivers, at this point, should damned well know their behavior is a problem. But they still routinely do it, day in and day out. There's no easy answer, and a lot of innocent people wind up paying for lack of one. That said, distracted driving doesn't mean we should be distracted (ugh) from other tasks like implementing a rigorous V2V solution that respects user privacy.

 

[McDeath]

Well, looks like someone else, maybe California, will have to establish rational standards now.

oh like how they were going to do with privacy but just listened to lobbyists instead?

Or how they were going to do healthcare, but just listened to lobbyists instead?

To be fair though, the lobbyists want this. Trump is just for slashing the government. But I wonder why business wants it.

I honestly highly doubt Trump is for anything concrete, except for doing the opposite of Obama. If Obama had rallied for reducing government programs, I could almost seriously see myself betting that Trump would try to expand same said programs.

 

[moridinga]

It actually makes more sense these days to come up with an industry driven (not government driven) standard. The government just needs to continue reserving the bandwidth slice already reserved for this, and let the vehicles companies hash out a standard. Right now that standard doesn't make sense because the technology is evolving so fast. Once driverless technology is more mature they can start looking at making a choice as to what v2v protocols are best suited for a mostly driverless world so we can start re-optimizing intersections to account for them.

Who do you think wrote the standards? I'll give you a hint. it wasn't the government. This is all about mandating adoption of the industry-written standards. The actually standards (things like IEEE 1609.4 and IEEE 802.11p) were not written by the government, nor by Trump, nor by a bunch of house elves. Technical standards come from technical bodies (like the SAE and IEEE).

 

[Kevin L. Nault]

Diligently reviewing comments to see if there's a better idea than your proposed rule?!? Shame on you, NHTSA! Adopt the Ajit Pai method of decision making and speed things up enormously!

1. Find the plan and implementation most harmful to the public your little brain can imagine;
2. Publish for comment;
3. If people actually comment, claim it was a malicious DDOS attack;
4. Ignore comments not handed to you by lobbyists with checks and/or job offers attached;
5. Do what you were going to do anyway.

Much more efficient, see?

 

[moridinga]

Good. V2V is troubling for privacy as it makes automated tracking even easier than plate scanning. It also won't work with motorcycles that tend to avoid these fancy features, bicycles, pedestrians, and other obstacles. As general obstacle avoidance technology is basically ready for production, is there even a need for V2V?

This is false, privacy has been taking into account as the signing keys are rotated after a short period of time. Vehicles are required to have like three years worth of keys with those keys having a lifetime of just a few minutes. Of course that means you periodically have to get new keys. The public key infrastructure surrounding the V2x effort is sort of bananas.

 

[kkeane]

Good. V2V is troubling for privacy as it makes automated tracking even easier than plate scanning. It also won't work with motorcycles that tend to avoid these fancy features, bicycles, pedestrians, and other obstacles. As general obstacle avoidance technology is basically ready for production, is there even a need for V2V?

here's the thing if vehicles have cellular modems in them then insurance companies can monitor everything you do with the vehicle and if you do one wrong thing ever then adios insurance.

Only if the insurance companies have access to that data. I don't see any requirement to give it to them on the horizon.

There doesn't need to be a legal requirement. All it takes is all the insurance companies putting it into their contracts.

 

[Dr Gitlin]

Update: Ars reached out to NHTSA this morning, which told us that it has yet to make a final decision. "The vehicle-to-vehicle notice of proposed rulemaking was released in December 2016 for public feedback, and received over 460 comments. NHTSA is still reviewing and considering all comments submitted and other relevant new information to inform its next steps. An update on these actions will be provided when a decision is made as part of the Department’s ongoing regulatory review," it told us in a written statement.

SO I would have liked this to have been at the top of the article...since it pretty much disproves the rest of the article so I could have just skipped it.

Fair point, I’ll make that change now.

 

[surfbot]

I'm seeing a lot of misinformation, so lets clear some things up. The proposed rule is public, so anyone can see what they are talking about.

Security-related problems are addressed with a strong crypto mechanisms using a newly standardized certificate format that enables PKI-based signing/authentication and encryption. V2V messaging under the proposed rule is limited to what is described in a few other standards that define message formats and performance requirements.

This is not about autonomous vehicles. This is effectively about creating a vehicle-based "here-I-am" transmission. Its a little more sophisticated than that, but that's the gist. Identifiers are not persistent, so tracking is difficult (though not impossible, tracking would require a footprint not dissimilar from similar approaches using LPR).

This technology is being deployed in various pilot projects around the United States, and also is included in a 2018 Cadillac direct from the manufacturer.

What is at risk? Well, given that industry and the federal government have spent ten years working out the kinks, identified safety benefits and done substantial hard ground truth testing, we have an idea what this technology could provide. No other technology has done more than one-off demonstrations or shout about how it might work. DSRC-based V2V has tons of hard research behind it including years of field tests with thousands of equipped vehicles.

Bottom line, its another sensor that functions in environments where even the newest vehicle sensors sometimes won't; fed into safety systems, this will save lives. Period. It is ready for deployment now, else those pilot projects wouldn't have any devices to buy.

This technology uses a swath of dedicated spectrum in the 5.9 GHz band. There are industry players that want that spectrum. Spectrum is valuable.

 

[THavoc]

I dunno. I've never been a big fan of V2V.

<<shudder>> I'm leaning towards agreeing w/ Trump on this one.

 

[kkeane]

Good. V2V is troubling for privacy as it makes automated tracking even easier than plate scanning. It also won't work with motorcycles that tend to avoid these fancy features, bicycles, pedestrians, and other obstacles. As general obstacle avoidance technology is basically ready for production, is there even a need for V2V?

This is false, privacy has been taking into account as the signing keys are rotated after a short period of time. Vehicles are required to have like three years worth of keys with those keys having a lifetime of just a few minutes. Of course that means you periodically have to get new keys. The public key infrastructure surrounding the V2x effort is sort of bananas.

That's just window dressing, though. Whether you have one key or a thousand, somebody - at least the car manufacturer, and thus the NSA and FBI, and probably every police agency in the country - has a list of all the keys for your car.

Actually, the complete list of keys has to be public - otherwise, there is no way to identify spoofed keys.

 

[notATroll123]

Confusing... I've seen articles here complain how V2V is a beurocratic morrass that by the time it is deployed will be obsolete. Yet it's bad Trump is pushing to cancel because..... It's Trump???... The guy may be an idiot but celebrate the wins where you can.

 

[kkeane]

I dunno. I've never been a big fan of V2V.

<<shudder>> I'm leaning towards agreeing w/ Trump on this one.

Surprisingly, Trump has done a number of good things when it comes to privacy (I don't give him credit for doing it *intentionally* but appreciate the effect nonetheless).

For one, he ended Operation Gatekeeper, the government program that used to deny legal but "undesirable" businesses access to such things as bank accounts. It was used against gun dealers and against (legal, first-Amendment-protected) pornography.

For another, it appears that he is dismantling the checkpoints along the freeway on the Mexican border. He probably did that so that he can claim that "arrests for illegal immigrants are down" but as a side effect, border patrol gets to do fewer warrantless searches of the cars of Americans.

And blocking V2V is another one in the same vein.

 

[notATroll123]

Good. V2V is troubling for privacy as it makes automated tracking even easier than plate scanning. It also won't work with motorcycles that tend to avoid these fancy features, bicycles, pedestrians, and other obstacles. As general obstacle avoidance technology is basically ready for production, is there even a need for V2V?

This is false, privacy has been taking into account as the signing keys are rotated after a short period of time. Vehicles are required to have like three years worth of keys with those keys having a lifetime of just a few minutes. Of course that means you periodically have to get new keys. The public key infrastructure surrounding the V2x effort is sort of bananas.

That's just window dressing, though. Whether you have one key or a thousand, somebody - at least the car manufacturer, and thus the NSA and FBI, and probably every police agency in the country - has a list of all the keys for your car.

Actually, the complete list of keys has to be public - otherwise, there is no way to identify spoofed keys.

But those lists don't have to associate the keys between each other...

Also couldn't the public portion of the key also be signed (or issued) by a well know entity like most of the internet https system works.

I'm sure I need to dive into the actual spec to look at this but technically there is no reason there "has to be" a master list.... Infact a master list would be a royal pain because how do you distribute and keep that updated with all cars ?

 

[rick*d]

"Hopes have long been pinned on V2V as a way to cut traffic fatalities, which have been on the rise the past two years. "

Traffic fatalities have been steadily rising ever since the introduction of texting on a smartphone which are now used by morons simultaneously while driving. THAT is one problem that really needs addressed ASAP.

That's a self-correcting problem. Survival of the fittest, Darwin awards, and all that.

 

[THavoc]

"Hopes have long been pinned on V2V as a way to cut traffic fatalities, which have been on the rise the past two years. "

Traffic fatalities have been steadily rising ever since the introduction of texting on a smartphone which are now used by morons simultaneously while driving. THAT is one problem that really needs addressed ASAP.

That's a self-correcting problem. Survival of the fittest, Darwin awards, and all that.

I don't know about that.

I can't count the number of times where I've read about a stupid (or drunk) person escaping serious injury or death but manages to kill other innocent people.

It's not as self-correcting as it should be.

 

[ziegler]

Update: Ars reached out to NHTSA this morning, which told us that it has yet to make a final decision. "The vehicle-to-vehicle notice of proposed rulemaking was released in December 2016 for public feedback, and received over 460 comments. NHTSA is still reviewing and considering all comments submitted and other relevant new information to inform its next steps. An update on these actions will be provided when a decision is made as part of the Department’s ongoing regulatory review," it told us in a written statement.

SO I would have liked this to have been at the top of the article...since it pretty much disproves the rest of the article so I could have just skipped it.

Fair point, I’ll make that change now.

Thank you. Damn Fine Job Sir. < insert thumbs up jesus >

 

[notATroll123]

Good. V2V is troubling for privacy as it makes automated tracking even easier than plate scanning. It also won't work with motorcycles that tend to avoid these fancy features, bicycles, pedestrians, and other obstacles. As general obstacle avoidance technology is basically ready for production, is there even a need for V2V?

here's the thing if vehicles have cellular modems in them then insurance companies can monitor everything you do with the vehicle and if you do one wrong thing ever then adios insurance.

Only if the insurance companies have access to that data. I don't see any requirement to give it to them on the horizon.

There doesn't need to be a legal requirement. All it takes is all the insurance companies putting it into their contracts.

If they did that adios to any customers.. All them having access to the data is allow for insurances to set risk groups more precisely(not getting into other privacy problems with them having it). If they boot customers and overcharge them someone else will recognise their mistake and gladly take their customers. Auto insurance does have some semblance of compitition.

 

[ziegler]

I dunno. I've never been a big fan of V2V.

<<shudder>> I'm leaning towards agreeing w/ Trump on this one.

Here's the thing. To help you deal with your guilt.

This was done in December 2016. There was a bunch of regulations that got pushed through between Nov7th and Jan 20th. Trump gave a blanket order to undo as many of them as was possible.

So you can console yourself that he is probably doing this just because Obama pushed them through during the "dead space" period more than anything technical/security/safety reason.

 

[ReaderBot]

Good. V2V is troubling for privacy as it makes automated tracking even easier than plate scanning. It also won't work with motorcycles that tend to avoid these fancy features, bicycles, pedestrians, and other obstacles. As general obstacle avoidance technology is basically ready for production, is there even a need for V2V?

here's the thing if vehicles have cellular modems in them then insurance companies can monitor everything you do with the vehicle and if you do one wrong thing ever then adios insurance.

Are you trying to argue that is V2V isn't killed, new vehicles won't have cellular radios?

 

[truthyboy15]

Diligently reviewing comments to see if there's a better idea than your proposed rule?!? Shame on you, NHTSA! Adopt the Ajit Pai method of decision making and speed things up enormously!

1. Find the plan and implementation most harmful to the public your little brain can imagine;
2. Publish for comment;
3. If people actually comment, claim it was a malicious DDOS attack;
4. Ignore comments not handed to you by lobbyists with checks and/or job offers attached;
5. Do what you were going to do anyway.

Much more efficient, see?

nah I guarantee they already made their decision they just don't want to announce anything or the grand czar has said glue your mouths shut.

 

[truthyboy15]

Good. V2V is troubling for privacy as it makes automated tracking even easier than plate scanning. It also won't work with motorcycles that tend to avoid these fancy features, bicycles, pedestrians, and other obstacles. As general obstacle avoidance technology is basically ready for production, is there even a need for V2V?

here's the thing if vehicles have cellular modems in them then insurance companies can monitor everything you do with the vehicle and if you do one wrong thing ever then adios insurance.

Are you trying to argue that is V2V isn't killed, new vehicles won't have cellular radios?

not at all just saying that with cellular modems in vehicles they will be used to hurt the customer and profit the business.

 

[tripodal]

Well, looks like someone else, maybe California, will have to establish rational standards now.

oh like how they were going to do with privacy but just listened to lobbyists instead?

Or how they were going to do healthcare, but just listened to lobbyists instead?

To be fair though, the lobbyists want this. Trump is just for slashing the government. But I wonder why business wants it.

As a driver, I don't want the liability or chance that someones else car instructs mine to do something i did not wish. My car, reacting to bad behavior is one thing, Your car telling me to slam on my breaks because yours is faulty is another.

not to mention the inevitable rooting of the control systems. "Install this app for a faster commute -- automatically instructs other cars to move over via the v2v system"

 

[powder]

the problem with V2V is unless you include strong signing or authentication, its going to be very hard to trust any data received, as there will always be the possibility that someone is faking data.

And there is no way in hell I'll willingly support something that is screaming my name/vin/number all down the street, its a 1984 surveillance persons wet dream.

I'm sorry, but you're talking out of your ass and spreading false information. You're not the first person to consider these issues. First, V2V includes a security credential management system with signed certificates and supports certificate revocation. Second, the onboard units do not broadcast unique identifiable information such as your name, VIN, or plate info.

The SAE J2735 Basic Safety Message, which is what your car broadcasts to everyone else, uses a random temporary identifier that changes frequently. That is, other vehicles can determine that a vehicle is at a certain position and trajectory but can't tell if it is the same vehicle as a minute ago. This also applies the MAC for the OBU device, which also cycles through random values.

 

[tripodal]

Good. V2V is troubling for privacy as it makes automated tracking even easier than plate scanning. It also won't work with motorcycles that tend to avoid these fancy features, bicycles, pedestrians, and other obstacles. As general obstacle avoidance technology is basically ready for production, is there even a need for V2V?

here's the thing if vehicles have cellular modems in them then insurance companies can monitor everything you do with the vehicle and if you do one wrong thing ever then adios insurance.

Only if the insurance companies have access to that data. I don't see any requirement to give it to them on the horizon.

There doesn't need to be a legal requirement. All it takes is all the insurance companies putting it into their contracts.

If they did that adios to any customers.. All them having access to the data is allow for insurances to set risk groups more precisely(not getting into other privacy problems with them having it). If they boot customers and overcharge them someone else will recognise their mistake and gladly take their customers. Auto insurance does have some semblance of compitition.

We all stopped using gmail when they started reading our emails for contextual adds, right?

 

[ProfessorGuy]

Good. V2V is troubling for privacy....

here's the thing if vehicles have cellular modems in them then insurance companies can monitor everything you do with the vehicle and if you do one wrong thing ever then adios insurance.

No, cancelling the policy is a money-loser. The correct play is to never monitor the connection, but save all the data forever, and keep collecting the premiums. Then when a claim is made, troll through the records to find the inevitable violations which, by the fine print, means they don't have to pay out.

 

[lowcrawler]

Now if they'll do away with mandatory backup cameras... that'd be nice.


(I'm not anti-backup-camera... I'm anti-mandating expensive solutions)

 

[TechfanMD]

"Hopes have long been pinned on V2V as a way to cut traffic fatalities, which have been on the rise the past two years. "

Traffic fatalities have been steadily rising ever since the introduction of texting on a smartphone which are now used by morons simultaneously while driving. THAT is one problem that really needs addressed ASAP.

We also had an economic recovery. People are driving more because they are working more and engaging in more leisure activities. More road miles means more accidents. Also, increase of congestion in commuter routes seems to have a more exponential effect on traffic accidents.

Generally statistics (for fatalities), when used for determining safety or impact of policy are normalized to be fatalities per mile driven or something similar - so it wouldn't matter if we're driving more or less, it is a metric of safety not total fatalities.

Edit: That said, they actually seem to be down for both number of deaths and number of deaths per miles driven. https://en.wikipedia.org/wiki/List_of_m ... S._by_year

 

[surfbot]

Good. V2V is troubling for privacy as it makes automated tracking even easier than plate scanning. It also won't work with motorcycles that tend to avoid these fancy features, bicycles, pedestrians, and other obstacles. As general obstacle avoidance technology is basically ready for production, is there even a need for V2V?

here's the thing if vehicles have cellular modems in them then insurance companies can monitor everything you do with the vehicle and if you do one wrong thing ever then adios insurance.

Are you trying to argue that is V2V isn't killed, new vehicles won't have cellular radios?

not at all just saying that with cellular modems in vehicles they will be used to hurt the customer and profit the business.

The peeps that want your car to have a cellular radio don't really want DSRC-based V2V. They want the DSRC spectrum for their own uses, and claim, despite a lack of real-world evidence, that emerging 3gpp standards can provide the same benefits. Meanwhile, crashes happen.

 

[rick*d]

Good. V2V is troubling for privacy as it makes automated tracking even easier than plate scanning. It also won't work with motorcycles that tend to avoid these fancy features, bicycles, pedestrians, and other obstacles. As general obstacle avoidance technology is basically ready for production, is there even a need for V2V?

here's the thing if vehicles have cellular modems in them then insurance companies can monitor everything you do with the vehicle and if you do one wrong thing ever then adios insurance.

Here's the thing: like OnStar, a significant percentage of drivers won't use the service beyond the "complementary" introduction period. I won't pay $35/month for LTE on a tablet; I'm certainly not going to subscribe for my car, and I'm sure I'm not alone.

V2V, on the other hand, is free, and much more likely to be used. The only reason V2V is dead is that the rent seekers can't make money off of it.

 

[markgo]

My feelings are mixed. The actual V2V protocol was ridiculously out of date, but if this means the reserved spectrum is gone...that's seriously bad for the hope of reliable car-to-car communication.

I strenuously disagree with this quote from the article:

In fact, it's going to be difficult to buy a new car in 2018 that doesn't have 4G LTE—which Nexar has shown is sufficient for short-range warnings—and 5G won't be far behind.

Unless someone manages to use LTE point-to-point with no cell tower, or at least ability to roam across ALL networks, there are huge swathes of highway that have poor (high latency) or no network connectivity for one or more networks. And these are where some of the worst weather occurs--high mountains.

Even in well connected areas--most people experience at least some short term network drops due to terrain or connecting to too distant tower. Would you want that to mess with real-time safety data communications?

The Nexar solution isn't even a general purpose tech. It looks like more crap targeted at getting insurance discounts by tracking your driving. Love how it uploads every sharp braking incident and claims that it's for your benefit.

Cell phone style connectivity is all fine and good for many connected car scenarios. But not real-time critical communications.

In a more perfect world, the government would just bless a network tech (say, a modified 802.11, just like the basis for the killed standard), and direct companies to use a open standard approach without licensing costs.

Since the government controls bandwidth, this isn't something that private parties can easily spin up on their own and competing standards work directly against safety, which demands interoperability.

But obviously, we're a long way from that perfect world.

 

[Akemi]

Well, looks like someone else, maybe California, will have to establish rational standards now.

oh like how they were going to do with privacy but just listened to lobbyists instead?

Or how they were going to do healthcare, but just listened to lobbyists instead?

To be fair though, the lobbyists want this. Trump is just for slashing the government. But I wonder why business wants it.

They think it will save on production costs. What they don't get is the first fully automated vehicle that causes an accident involving serious injury or death is going to net a gargantuan lawsuit and settlement. It's also likely to spur a bout of public outrage and pressure for over-regulation. This stance is myopic.

 

[Maury Markowitz]

> to warn each other of unseen hazards

Just this morning I heard on the radio that an accident occurred near my childhood home. Traffic was slowing down around 8 PM the night before when a semi failed to notice and slammed into the back of the queue. A huge pile up resulted in front as the cars were jammed together, trapping people inside. He was carrying gasoline, which spilled onto the roadway and ignited. Dozens of cars burned, melting down into metal puddles.

This keeps happening. About 15 years ago a fog blew across the 401 on the Detroit end, hundreds of cars piled up and then caught on fire, dozens dead. Last year they were doing construction on the 401 just north of my current home. A semi piled into traffic slowed for the construction. Many dead. A week or so later the same thing happened on the other end of the same construction.

I, for one, want something like V2V decades ago.

 

[THavoc]

> to warn each other of unseen hazards

Just this morning I heard on the radio that an accident occurred near my childhood home. Traffic was slowing down around 8 PM the night before when a semi failed to notice and slammed into the back of the queue. A huge pile up resulted in front as the cars were jammed together, trapping people inside. He was carrying gasoline, which spilled onto the roadway and ignited. Dozens of cars burned, melting down into metal puddles.

This keeps happening. About 15 years ago a fog blew across the 401 on the Detroit end, hundreds of cars piled up and then caught on fire, dozens dead. Last year they were doing construction on the 401 just north of my current home. A semi piled into traffic slowed for the construction. Many dead. A week or so later the same thing happened on the other end of the same construction.

I, for one, want something like V2V decades ago.

But the same thing can be done w/ the auto-brake systems that are being put in now.

There's no need to share or interconnect information like V2V is suppose to do.

 

[KKSka]

Hey, even a broken clock is right twice a day. In this instance, the Trump administration is right (assuming that's what NHTSA ultimately decides) to go technology agnostic. This is an instance where the federal bureaucracy can't keep up with technological change.

 

[Bluestrike2]

I'll just reply to two comments at once since there's an overlap between them:

the problem with V2V is unless you include strong signing or authentication, its going to be very hard to trust any data received, as there will always be the possibility that someone is faking data.

And there is no way in hell I'll willingly support something that is screaming my name/vin/number all down the street, its a 1984 surveillance persons wet dream.

On the first point, the potential danger is fairly well-contained. V2V data wouldn't replace information gleened from a car's own (trusted) sensors. Its eventual rollout will be gradual given that there are ~254 million cars on US roads today with an average age of 11.4 years. V2V data has to be treated more along the lines of secondary warning information than anything definitive. In other words, V2V data can only be used to better inform a car's decisions in a very specific context. It can't be used to make those decisions outright because the car *has* to assume that in a given scenario there's no data to be received.

That takes care of both the problem of old, non-connected ("dumb") cars and potentially malicious actors. Even if they're transmitting false data, it's not relied upon. An autonomous car won't be tricked in that matter. A manually-driven car might pop up nuisance alerts to the driver, but even with manual cars we're seeing more sensors added with each new model year. The most likely attack vector would be a DoS attack; but even if they overwhelm V2V communications in an area, it won't result in mass crashes.

That could change somewhat if there were some sort of cheap, mandated retrofit kit that plugged into a car's OBD-II port, or in the case of older cars without OBD-II, had some internal sensors of its own that could be used to broadcast basic "here I am, and this is how fast I'm going" data. Dumb, one-way communication that's slightly less "dumb" than no communication at all.

In any case, the "useful, but verify" approach also protects against malicious actors. Because V2V data isn't used as the sole criteria for autonomous/semi-autonomous decisions (and it's moot issue for manually-driven cars), hackers aren't going to be in a position to start crashing cars en masse. At least not through V2V itself. And the same approach applies to V2I communication with fixed road infrastructure.

On the second point, why would V2V systems have to broadcast plate numbers, car VINs, or your personal contact information? None of that information is relevant to V2V's purpose. That concern is handled easily enough by telling your legislators to exclude it from any legal standards that are eventually written.

Well, looks like someone else, maybe California, will have to establish rational standards now.

oh like how they were going to do with privacy but just listened to lobbyists instead?

Or how they were going to do healthcare, but just listened to lobbyists instead?

To be fair though, the lobbyists want this. Trump is just for slashing the government. But I wonder why business wants it.

As a driver, I don't want the liability or chance that someones else car instructs mine to do something i did not wish. My car, reacting to bad behavior is one thing, Your car telling me to slam on my breaks because yours is faulty is another.

not to mention the inevitable rooting of the control systems. "Install this app for a faster commute -- automatically instructs other cars to move over via the v2v system"

V2V communications don't cede control to external systems or data provided by the same external sources. Nor would V2V be the only available source of information about other cars on the road. Even in non-autonomous cars, we're seeing additional sensors being added for everything from adaptive cruise control to self-braking systems and collision avoidance systems. V2V is another data point for those systems. There's no centralized (or decentralized) traffic control system in play. The car still makes its own decisions. It just has more data it can use to get a better sense of the road beyond its own sensors. In the far future, if there is eventually some sort of centralized traffic control, it'll no longer be the same V2V that we're talking about.

There are plenty of potential problems to consider with V2V, particularly with driver privacy. But some of the concerns in the comments aren't really applicable.

 

[rick*d]

Good. V2V is troubling for privacy as it makes automated tracking even easier than plate scanning. It also won't work with motorcycles that tend to avoid these fancy features, bicycles, pedestrians, and other obstacles. As general obstacle avoidance technology is basically ready for production, is there even a need for V2V?

here's the thing if vehicles have cellular modems in them then insurance companies can monitor everything you do with the vehicle and if you do one wrong thing ever then adios insurance.

Are you trying to argue that is V2V isn't killed, new vehicles won't have cellular radios?

They won't have cellular radios constantly broadcasting your identity, position, and speed. V2V would do that, so cellular wouldn't have to. Cellular would be restricted to entertainment, navigation, etc. Unlike cellular, V2V doesn't have the range for insurance companies to constantly monitor it.

 

[rick*d]

Hey, even a broken clock is right twice a day. In this instance, the Trump administration is right (assuming that's what NHTSA ultimately decides) to go technology agnostic. This is an instance where the federal bureaucracy can't keep up with technological change.

You mean like when the Reagan Administration decided not to pick a standard for stereo AM radio and let the market choose? That worked out real well, didn't it? I mean, look at all the stereo AM radios we have on the market today.