Vehicle-to-Everything Technology
A regional pilot program could be expanding.
AP (“Cars collect troves of data about traffic and road hazards. Should they share it?“):
The secret to avoiding red lights during rush hour in Utah’s largest city might be as simple as following a bus.
Transportation officials have spent the past few years refining a system in which radio transmitters inside commuter buses talk directly to the traffic signals in the Salt Lake City area, requesting a few extra seconds of green when they approach.
Congestion on these so-called smart streets is already noticeably smoother, but it’s just a small preview of the high-tech upgrades that could be coming soon to roads across Utah and ultimately across the U.S.
Buoyed by a $20 million federal grant and an ambitious calling to “Connect the West,” the goal is to ensure every vehicle in Utah, as well as neighboring Colorado and Wyoming, can eventually communicate with one another and the roadside infrastructure about congestion, accidents, road hazards and weather conditions.
With that knowledge, drivers can instantly know they should take another route, bypassing the need for a human to manually send an alert to an electronic street sign or the mapping apps found on cellphones.
“A vehicle can tell us a lot about what’s going on in the roadway,” said Blaine Leonard, a transportation technology engineer at the Utah Department of Transportation. “Maybe it braked really hard, or the windshield wipers are on, or the wheels are slipping. The car anonymously broadcasts to us that blip of data 10 times a second, giving us a constant stream of information.”
So far, so good.
I have, for quite some time now, routinely used various mapping apps even for routes I know well—including if not especially my daily commute—in order to be alerted to traffic conditions and possible route alterations. Increasing the amount of data collected to that end is an unalloyed good.
A 2016 analysis by the National Highway Traffic Safety Administration concluded V2X could help. Implementing just two of the earliest vehicle-to-everything applications nationwide would prevent 439,000 to 615,000 crashes and save 987 to 1,366 lives, its research found.
Dan Langenkamp has been lobbying for road safety improvements since his wife Sarah Langenkamp, a U.S. diplomat, was killed by a truck while biking in Maryland in 2022. Joining officials at the news conference announcing the vehicle-to-everything blueprint, Langenkamp urged governments across the U.S. to roll out the technology as widely and quickly as possible.
“How can we as government officials, as manufacturers, and just as Americans not push this technology forward as fast as we possibly can, knowing that we have the power to rescue ourselves from this disaster, this crisis on our roads,” he said.
Again, this makes sense. The data is being collected; it may as well be put to good use.
Who it’s shared with and for what purposes, obviously, is the rub.
Most of the public resistance has been about privacy. Although the V2X rollout plan commits to safeguarding personal information, some privacy advocates remain skeptical.
Critics say that while the system may not track specific vehicles, it can compile enough identifying characteristics — even something as seemingly innocuous as tire pressure levels — that it wouldn’t take too much work to figure out who is behind the wheel and where they are going.
“Once you get enough unique information, you can reasonably say the car that drives down this street at this time that has this particular weight class probably belongs to the mayor,” said Cliff Braun, associate director of technology, policy and research for the Electronic Frontier Foundation, which advocates for digital privacy.
I’m honestly more concerned about this information being routinely fed to law enforcement, insurance companies, and others, which seems a more likely target for privacy concerns than one-off tracking of the mayor’s, much less my, vehicle. Indeed, some insurance companies already encourage customers to use automatic vehicle monitoring technology—ostensibly, to get better rates. It would certainly be a massive revenue stream for police departments to be able to automatically ticket any vehicle for any moving violation—which the driver would be unable to defend against since the notification would likely happen well after the event.
The federal blueprint says the nation’s top 75 metropolitan areas should aspire to have at least 25% of their signalized intersections equipped with the technology by 2028, along with higher milestones in subsequent years. With its fast start, the Salt Lake City area already has surpassed 20%.
Of course, upgrading the signals is the relatively easy part. The most important data comes from the cars themselves. While most new ones have connected features, they don’t all work the same way.
Before embarking on the “Connect the West” plan, Utah officials tested what they call the nation’s first radio-based, connected vehicle technology, using only the data supplied by fleet vehicles such as buses and snow plows. One early pilot program upgraded the bus route on a busy stretch of Redwood Road, and it isn’t just the bus riders who have noticed a difference.
“Whatever they’re doing is working,” said Jenny Duenas, assistant director of nearby Panda Child Care, where 80 children between 6 weeks and 12 years old are enrolled. “We haven’t seen traffic for a while. We have to transport our kiddos out of here, so when it’s a lot freer, it’s a lot easier to get out of the daycare.”
That simply changing the behavior of fleet vehicles would have that much impact—in Utah, no less—is counterintuitive.
Casey Brock, bus communications supervisor for the Utah Transit Authority, said most of the changes might not be noticeable to drivers. However, even shaving a few seconds off a bus route can dramatically reduce congestion while improving safety, he said.
“From a commuter standpoint it may be, ‘Oh, I had a good traffic day,’” Brock said. “They don’t have to know all the mechanisms going on behind the scenes.”
Traffic congestion is a phenomenon that I still don’t pretend to understand. Much of it is attributable to “phantom traffic jams”:
And human nature being what it is, no amount of tips about not weaving in and out of traffic, not tailgating, and the like is going to help. There are just too many moving parts. So, automating signals and the like to be attuned to real-time traffic patterns would help tremendously.
Cavnue, a Washington, D.C.-based subsidiary of Alphabet’s Sidewalk Infrastructure partners, funded the Michigan project and was awarded a contract to develop the one in Texas. The company has set a goal of becoming an industry leader in smart roads technology.
Chris Armstrong, Cavnue’s vice president of product, calls V2X “a digital seatbelt for the car” but says it only works if cars and roadside infrastructure can communicate seamlessly with one another.
“Instead of speaking 50 different languages, overnight we’d like to all speak the same language,” he said.
I haven’t the foggiest what all of this costs. But it’s the type of infrastructure investment that makes sense to federalize for standardization and economy of scale.
1) no one wants to pay for the V2X infrastructure and to maintain it. Hence why more and more gets crammed into the cars.
2) it’s much easier if everything is V2V at the same time. A sizable percentage of the patent applications I see are addressed specifically to the problem of traffic control of “dumb cars” driving together with “smart” cars.
3) too many applications I see work on the Underpants Gnomes theory of planning, with “add AI” taking the place of “stealing underwear”.
There’s a Mythbusters ep where they tested that. It was that which let me understand the phenomenon. I’ll try to hunt it down later. As I recall, they set a number of cars to drive around in a defined circular track and had one break or reduce speed. You can see the slowdown spread around the track.
I use Waze daily on the drive to and from work. I tend to stick to one route, unless the estimated time difference is big, say over 5-10 minutes. In particular when it suggests a circuitous route that adds like 2-5 kilometers. I won’t take such a detour to save maybe 3 minutes, especially since the traffic may deteriorate while I’m 3 kilometers into a detour.
In the past, I’ve seen several well thought through articles that contend that the only way autonomous cars will work is if the vehicles themselves communicate with each other. This would require a far more robust internet than we have now or at least some of bluetooth-like communications that would automatically connect, when the vehicles sensors detect another vehicle nearby.
For vehicles and even pedestrians that lack any type of V2V system or autonomous sensors, there could be a phone app that alerts vehicles of presence. The problem now is that most of the effort going into use of this data is in how to monetize it.
All new time-saving and money-saving technologies over promise and underperform.
@Sleeping Dog: the problem is that the more bells and whistles you put in the vehicle, the more expensive it will be and the more of a push towards attempts by auto companies to insist on rolling subscriptions for everything.
( one reason why I’d never get a Tesla. Chances are too high that Musk will get bored with keeping the software updates coming, suddenly say that Tesla is getting out of electric vehicles entirely, and you discover the next morning your car has been bricked.)
@Grumpy Realist:
I’ve never seriously considered a Tesla, even when I was buying an EV. I don’t like a manufacturer that starts out by lying to me, and no one in the automotive sector lies like Elon. Elon is unstable and a fascist, BMW is neither. Also Tesla interiors are as charmless as a hotel business center. And there’s the cluelessness of killing steering wheel controls in favor of a touch-screen. And the build quality. And the customer experience. So many good reasons not to give Elon a dollar.
@Grumpy Realist:
The other day I ran into a video claiming getting a Cybertruck through a car wash bricked it. I hunted down details, as the video made claims without providing much evidence. The more nuanced story is the owner did put his Texla Cybertruck through a car wash, then drove home, and then the car’s systems stopped working. He tried a reboot, and nothing happened. So he sent either a message or email to Texla support.
The next day the weird truck was back to normal. Support did contact him, and said the truck required an update and reboot, which took the totally brief and reasonable time of 5 hours.
Imagine you need to get somewhere urgent, and your vehicle is indisposed while it updates and reboots. I’d never get a car that might fail that particular way.
And I would’t buy a Texla because the money would go to Xlon the Fascist.
Drove to Boston and back this weekend to see daughter. We had multiple slow downs and times when traffic actually stopped with no apparent reason, no accidents and no construction ie phantom traffic jams. Link goes to a description, why it occurs and a possible remedy. A lot of it is due to the tailgating common on highways but it’s made worse by high volumes. Looks like with just beer spacing while driving we could increase highway volumes by 50% and avoid slowdowns. Article suggests this could be done with some modifications to adaptive cruise control systems but I doubt it works without autonomous cars.
I have thought for a while self driving cars will be resisted because lots of people, like me, routinely drive faster than they should, 7-10 mph over the posted limits. However, there are always people in the fast lane who arent driving fast. So the reality is that people like me probably spend 2/3 of our time driving like we want and the other 1/3 driving much slower. At least for me I would gladly give up that extra time of driving faster in exchange for the lack of irritation at the slow drivers living in the left lane.
https://www.livescience.com/61862-why-phantom-traffic-jams-happen.html
Steve
@Grumpy Realist:
Today pretty much, every new car has sensors that detect another vehicle in proximity and have rudimentary ability to drive themselves, under supervision. Adding a bit of communications software and a few hardware bits, will be far cheaper than engineering a truly fully autonomous car that can operate in most conditions. If all cars broadcast their location and movement, the tech requirement to accomplish autonomy would go down.
Google maps, Waze, and other apps determine the state of traffic by tracking the location and speed of cell phones. But which phones? All the phones in the network with location turned on, or just those running a certain app? Since Google owns Waze, we can assume it shared data with Google maps. Past that, no clue.
Waze sometimes gets it wrong. Lots of time it marks a solid red line indicating heavy traffic. Sometimes it even gives the voice warning “heavy traffic reported ahead.” On many such occasions, the street is literally empty for several blocks marked as having heavy traffic. This happens so often in one of my routes home, that I wonder whether lots of people in the office buildings on that street are running Waze or Maps at their desks for some reason (not likely).
Materially it doesn’t matter much. but ti’s those kinds of obvious glitches that make me nervous to automate cars for self driving.
Phantom / shockwave traffic jams are pretty easy to “simulate” and you can find videos on youtube of it happening if you search those key words.
Ford has already patented an extensive setup that uses your car’s infortainment system to both play ads and to record all conversations for AI parsing.
https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/20240289844
Yup in the very near future you’ll get blasted with ads in and out of your car while anything you say that might remotely be criminal regardless of sarcasm or context is dutifully reported to the authorities. Instant tickets ala demolition man for not stopping long enough or driving slightly too fast etc.
Hackers gonna have a field day with this shit. It’s going to be the KIA/Hyundai/mopar vulnerabilities fleet wide.
@Sleeping Dog:
Until a few bored people hack the system and use it to send invalid data to cause chaos or force other cars to move out of their way.
Are they planning to use the existing 5g systems for this or encroach further into publicly available frequencies?
I wouldn’t worry about the insurance companies ratting out their customers to LE. First off, that would be a form of sales-suicide. Word would get out. Second, what the insurance company apps look for are sudden deceleration, speed of over 80 mph, and driving at “high risk” times, meaning late at night. That’s it. They would have to keep close track of speed limits for each and every street to be of use to LE, which is a lot of work and actually unnecessary work for their purposes.
Their actuaries were rather clever about sudden deceleration being key IMO. If you get tagged for one you can call and have it expunged from your record, something like 3-4 times a year they will do that but no more. It seems having lot of those is a very strong indicator of poor driving habits.
@Matt: IoT (Internet of Things) is well known for never updating, and being full of security holes. It’s not much of a problem now, because hacking a refrigerator or air purifier doesn’t give you much benefit.
All the cars on the roads, and traffic signals (and dynamic speed limits, I expect)? There’s lots of mischief/terrorism that could make use of that. At the most innocuous, I would like fewer red lights for me.
@dazedandconfused: You just unwittingly made the argument that libertarians make about how the free market will keep companies in line. How no company would ever poison their users because word would get out and they’d lose customers…
It’s not the insurance companies that would be doing the ‘ratting’. It would be the automotive maker’s AI that would do the ratting. You’d be lucky to even find out the source since it’s buried under so many lawyers and laws. Since this data is being sold to anyone one and everyone law enforcement can snatch it up with no oversight.
Speed limits for all streets are already tracked by LE. All you have to do is use a little AI to overlay the GPS coordinates/speed over known speed limits. Hell my infotainment GPS system is almost always correct on speed limits despite being +4 years out of date.
@Gustopher: You say that but IOT devices are regularly used by hackers for a variety of uses from mail spam to DDOS attacks. Zombies can make you a lot of money if you have the right people to sell access to. Anything with a decent amount of local storage can be used to hold illegal material.
Remember this oldie but goodie?
https://www.bbc.com/news/technology-25780908
I would like to include in my post about my GPS speed limit being correct the vast majority of the time. I’ve driven in 9 different states so far with that car so it’s not like I’m only puttering around one area.
Google maps can be helpful… Until it keeps trying to route me through a closed on-ramp.
Anyway… Highways should be the first, easy win for self driving car, as long as they can physically segregate them from manually driven cars. Don’t try to mix humans and automation. That doesn’t lead to happy times.
@Argon: my initial experience with navigation devices was Not Good when the bloody instructions tried to drive me through a construction site.
I’m still hesitant about Google Maps in the vicinity of tollbooths. Don’t listen to any screaming about what lane you need to be in until you get back into a section with the normal number of lanes.