US feds want cars conversing by 2017 New US cars and light trucks will soon be equipped with crash-avoidance vehicle-to-vehicle (V2V) communications systems, says the US Department of Transportation's National Highway Traffic Safety Administration (NHTSA). "Vehicle-to-vehicle technology represents the next generation of auto safety improvements, building on the …
Also, your position and direction and speed are public knowledge already (to anyone who uses their eyes) so broadcasting this info digitally is no big deal.
Yes, but with a persistent (registered) ID broadcast along with your speed, the police would only need to set up 'listening posts': no need for speed cameras, just log the data flow and file the tickets.
That's the trick about "expectation of privacy in public" - Anybody can observe what you are doing, it gets creepy when they record what you are doing. It gets very creepy when they follow you around all day notating your activities at all times.
Yes, but with a persistent (registered) ID broadcast along with your speed, the police would only need to set up 'listening posts': no need for speed cameras, just log the data flow and file the tickets
Did I say use a persistent ID? No I didn't. In fact you can tell by the word "Also" that I was adding to what was already said by the previous poster, who specifically said there should NOT be a persistent ID.
Can I make this clear to all the downvoters: I DON'T WANT A PERSISTENT ID!!!! Got that now? Good.
Or a 64 bit ID should be enough for a few million years of all cars switching IDs every second.
It is already mandated by law that all cars have number plates which can be used to track you. There are brake lights that show your intent to slow down. There are indicators on all non Audi cars that are legally mandated to indicate intended direction changes.
There is no premise of privacy with the sort of data needed by this system. As long as it goes peer to peer in a short range rather than over some cloud, what is the complaint?
Why is there any need for an ID at all if the goal is to alert the driver to a possible collision? If the car gets signals with position/speed/direction data (plus maybe vehicle type such as private/van/light truck/18-wheeler) from the vehicles in the immediate vicinity it should be able to create a representation of what is really happening. No need for an ID, no need to keep the info, etc.... Anything else has purposes other than safety, IMHO.
Perhaps I'm missing the point. If the purpose is to have vehicles talk to adjacent vehicles then it seems redundant since there are already various sensors that are perfectly capable of measuring and/or calculating that information. Either the system is relaying critical information from several vehicles ahead in order to provide a maximum reaction window or it seems a bit pointless. It would also enable some miscreant to broadcast false data hoping to cause an accident for whatever reason.
Perhaps I'm missing the point. If the purpose is to have vehicles talk to adjacent vehicles then it seems redundant since there are already various sensors that are perfectly capable of measuring and/or calculating that information. Either the system is relaying critical information from several vehicles ahead in order to provide a maximum reaction window or it seems a bit pointless.
Yes the system will relay information from several vehicles ahead. This is useful if normal visibility is obscured by, say, a huge truck, or fog. Of course people *should* be driving to the conditions, for example leaving adequate stopping distance between themselves and the truck in front and slowing down in fog, but we all know that in reality many people don't.
"There are indicators on all non Audi cars that are legally mandated to indicate intended direction changes."
On Audis these have been repurposed to all flash together. They allow the Audi owner to legally park on double yellow lines and in disabled spaces, as long as they are just popping in for a minute.
IMHO, the responses to my comment have been a reaction to the collection of data. That is a fair enough reaction to have, but I don't think it addresses the challenge I put out there.
The argument that an ID is unnecessary is wrong. I am not saying that a persistent ID is mandatory, nor that that ID needs to be linked in any way to your identity, but if your car one moment receives a message that an oncoming car has veered into your lane, loses that signal as you go down a hill and then you get another message saying an oncoming car is not in your lane, is it the same car? Being able to trace and then forget other cars for the "session" is important even if the ID number is randomly generated at each engine startup.
These "listening posts" as argued are no more invasive than a suggestion that the state could put out cameras with numberplate recognition software that works out the vehicles speed (either by radar or by listening to these "anonymous" messages and then taking a picture). Both need to be regulated to ensure a good balance between privacy and safety.
I am simply pointing out that from a privacy perspective, your car is already traceable with a camera and the CPU power of a Pi using off the shelf software. How do you think Google maps blurs the plates if it can't locate them with very simplistic pattern recognition?
Based on the reporting, I'm not seeing the benefits. Most (not all) of the accidents which are filed under "driver inattention" aren't going to be alleviated by this gadget - by the time the driver looks up from their smartphone/book/whatever and realises what is happening, the accident is bound to happen.
This is *not* a collision avoidance/reduction device - this is a device that the insurance companies can use to tell you that you had in fact been warned and thus the accident is your fault and we won't be paying thank you very much.
There are two ways to develop a driverless car. Either try to build the entire system in one go (what Google are doing) or build lots of small separate systems one by one which will eventually, when all working together at the same time, provide total control (what many major car manufacturers are doing).
Once this system is developed and working it can easily be linked to the existing collision avoidance system already used in many modern cars to improve that system, plus it can help your car choose a safe following distance.
Want to pester someone whose politics don't suit the New Order? Hack their sensors to feed them inaccurate warnings of traffic dangers. Or find a way to remotely have them break down at various inopportune times.
(Sad that this kind of crap actually has to be considered. Thanks, NSA fu@#ers!)
My knee-jerk reaction is that only on-board sensors can be relied upon to affect automatic actions such as braking. One cannot rely on an external signal (V2V) at all. Even just alerting the driver through V2V may be questionable - someone may decide to spoof the signal to spook or just annoy a driver. Besides, once on-board sensors are installed there will be no need for V2V for collision prevention at all. (V2V+relaying may improve traffic flow, provide information on traffic conditions, etc., stuff Waze does for you today in exchange for personal identification / location info / snooping / ads / etc.)
What everybody seems to have missed is that an active V2V system is pretty much useless until a significant fraction of vehicles are equipped with it.
In the gliding world we have FLARM, a short-range active GPS-based system in which every set broadcasts its 3D position and velocity vector while using the data sent by other sets to determine whether a collision is likely. Virtually every glider, helicopter and light plane operating in the Alps are now equipped and elsewhere in Europe coverage is, I believe pushing up to 50%. Experience has shown that FLARM was pretty much useless when less than 25% of local gliders carried it. Now it is starting to become worthwhile as usage exceeds 50% and so the remaining parts of the fleet are seeing that carrying it is a positive benefit and installing it too.
I'd say that V2V is less use on the road because a driver's traffic scan only has to cover the horizontal plane and in any case road vehicles already carry conspicuity features (lights, horns) and intentional signalling equipment (turn indicators and brake lights). Now add in the fact that FLARM is a small, self-contained box the size of a mobile phone that's easy to install in almost any cockpit while a road V2V system will need both something on the dash and (probably) external front and rear sensors. As a result V2V retrofit would probably not be easy or cheap. Consequently, V2V installation is likely to only be a feature of new vehicles and will be opposed on cost grounds by many owners.
The fleet coverage statistics for gliders are likely to apply to road users as well so, if we assume that a hypothetical V2V system is only available as an optional extra on new vehicles, how long is it likely to take for over half of all vehicles to be fitted with V2V sets?
To assume that you or your sensors will be able to spot every approaching danger on the road seems a bit short sighted. Things happening a few cars in front or behind could impact you before you have time to react, especially if visibility is reduced. If nothing else this V2V stuff could reduce the number of cars that find themselves slamming into the back of a multi-car pile-up.
It's also possible to "hack" the brake lights so they don't come on when you slow down. You have to weigh up the probabilities against the consequences and decide whether the safety features made possible by short range v2v conversations which may be beneficial where something happens out of line of sight (like just beyond the crest of a hill or a blind corner)
Surely the collision avoidance (braking) systems being fitted to cars now are far better than just a buzzer that will drive people mad in built up areas.
The car 100 metres (sorry yards) ahead is stationary, brake, stop, avoid, oh god your going to die!. Oh sorry it was in a parking bay on the side of the road.
Says pretty much all you need to know about this particular "cunning plan."
I suppose we should be grateful they did not feel the need for some kind of central reporting system to give a kind of nationwide radar picture (technically a Plan Position Indicator display) of the whole country.
Still there's always V 2.0.
For things like planes or ships which tend to travel in straight lines and normally pass each other with large separation collision prediction is practical.
Roads are not straight, the difference between normal passing and a collision is sometimes only a few feet. I can't see how it can give valid warnings sufficiently early without also giving lots of false warnings.
Also it can't be reliable till everything on the road is fitted with it and drivers relying on it before it is reliable will probably do more damage than is saves.
Looks like an expensive technological non-solution to the problem. The biggest benefit will be to those making money from it directly and from back handers.
This sort of crap is why I pay big bucks into my American Motorcycle Association "legislative action" fund that directly pays for lobbyists to fight it. I'm putting my money where my mouth is.
We're also fighting a requirement for "event data recorders" (aka "black boxes") and the EPA push for 15% ethanol gas.
We've managed to defeat Florida's law requiring ethanol in gas, and I no longer have to pay under the table for ethanol-free gas for my lawnmower and other things.
The EDR battle is lost. Manufacturers have been installing them for years and one forensic engineering company has a list (PDF) of models their data reader supports.
Just have a nice proper device that tells other vehicles some seriously bad information. Now one can be a pedestrian on an overpass of a busy freeway (I-280 nearby) and watch the cars. Maybe they can have contests to see of one can lite up brake lights in a character pattern.
Of course, if one is in Southern California, the multitude of lanes on I-210 might make a nicer bit-mapped display (they have 5 lanes there).
I welcome these types of moves - eventually it will lead to self driving cars. I figure that by the time I'm elderly, self driving cars will be well proven and affordable. And I'll be able to get around safely, when otherwise I'd be a danger. Also there would no longer be drunk drivers on the road - they would just be drunk passengers.
Have you ever driven with a nervous passenger that gasps and grabs for a hand hold all the time. It drives me nuts as I immediately have a mini-panic and look for the lorry bearing down on me. The last thing I want is a beeper or voice warning to replace this nervous-nelly. This is like the useless daytime driving lights. A solution looking for a problem.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
