Autonomous Driving Idea

[Max*]

Vehicle to Vehicle and Vehicle to Infrastructure [V2I] (stop lights, signs, etc) are already ideas that have been in the works for years. Two problems with your idea: as other mentioned you can't keep the non-autonomous cars out of the lane and even if you did then you are simply increasing congestion on already congested motorways. Second, 100mph would be inefficient energy wise, might be better with such a low drag coefficient [Model 3] than other cars though.

Autonomous cars shouldn't need special lanes. If done properly they should be able to integrate anywhere and adapt to unseen conditions.

Just preventing crashes would be enough to alleviate some congestion at least in my city.. One accident can slow down thousands of drivers.

Sure you can.

First of all, I assume he's referring to doing this on a highway. So pedestrians, stop signs, etc. don't count.

Second of all, where I live we have a 6-lane highway. 4 lanes are free (55mph). 2 lanes are EZ-pass only (65mph), with special on-ramps and off-ramps, with a variable $/mile rate depending on traffic flow in the free lanes.

Here's a visual with the ez-pass toll readers. The white things are dividers, you can't get in-out of it like traditional EZ-pass lanes: Google Maps

So you can implement something like this during the transition between autonomous and non-autonomous cars. Where you have (on large highways) an extra lane for autonomous only vehicles (think HOV only), which would encourage the adoption of autonomous vehicles, and help reduce congestion due to the higher speed limit.

 

[butchmank9]

By you came up with it, do you mean you read it somewhere a while back, then forgot about it, then came up with it again thinking it's your own?

This idea has been out for years, and has even been posted/discussed somewhere on TMC.

If I copy right someone idea then you can't just tell that person there idea is are ready out there show them! A copy right idea would have to be pretty close to what they wrote about. Taking away the HOV, putting a sensor in the lane. Was all of this mentioned in the article? This is just a helpful idea. i'm not trying to take away from anybody else's. Even if it was out their years ago when autonomous cars was not even thought about like today it is now becoming a reality is a more suitable time frame.

 

[JeffK]

Second of all, where I live we have a 6-lane highway. 4 lanes are free (55mph). 2 lanes are EZ-pass only (65mph), with special on-ramps and off-ramps, with a variable $/mile rate depending on traffic flow in the free lanes.

That would costs millions in infrastructure when it's completely unnecessary if autonomous cars work properly. What you have there is just a money grab.

 

[Max*]

That would costs millions in infrastructure when it's completely unnecessary if autonomous cars work properly. What you have there is just a money grab.

I'm just giving an alternative to not having regular cars in autonomous lanes. It's possible, I never said it'll be cheap.

 

[dsvick]

The obvious solution is to make them autonomous and flying!

 

[Saghost]

Still need a way to prevent non-autonomous vehicles from pulling into that lane to avoid the congestion in the other lanes, which already occurs today with non-qualifying vehicles using the HOV lane. In the case of the proposed autonomous lane, these cheating cars will also pose a safety risk in addition to just breaking the rules.

Cheating cars shouldn't pose a safety risk - the self driving swarm can see them just as well as they see other obstacles. A couple of cheating cars won't even have a huge impact on speed/flow, since they won't be able to set up the feedback loops that lead to stop and go traffic waves.

Having said that, cheating cars are certainly a problem that needs to be addressed.

Not necessarily a hard problem, though - any legitimate cars in the lane will be talking to the rest of the swarm, so every car in the swarm will know who cheaters are, and every car has both cameras and telecommunications.

It should be fairly trivial to have the car behind the cheater automatically upload a video and contact nearby patrol cars...

 

[NikeWings]

Cheating cars shouldn't pose a safety risk - the self driving swarm can see them just as well as they see other obstacles. A couple of cheating cars won't even have a huge impact on speed/flow, since they won't be able to set up the feedback loops that lead to stop and go traffic waves.

Having said that, cheating cars are certainly a problem that needs to be addressed.

Not necessarily a hard problem, though - any legitimate cars in the lane will be talking to the rest of the swarm, so every car in the swarm will know who cheaters are, and every car has both cameras and telecommunications.

It should be fairly trivial to have the car behind the cheater automatically upload a video and contact nearby patrol cars...

Not sure AP is quite ready yet for a car pulling into the HOV lane from a stopped position at 5mph, while the swarm is traveling 80 mph. Idiots are a graver concern than cheaters.

 

[Saghost]

Not sure AP is quite ready yet for a car pulling into the HOV lane from a stopped position at 5mph, while the swarm is traveling 80 mph. Idiots are a graver concern than cheaters.

Not much you can do about idiots like that, except physically separate/block the lane, as some current HOV lanes are.

The swarm cars are subject to the laws of physics, and your idiot might be setting up an unavoidable accident - but if he isn't, the swarm has a better chance of not having an accident than anyone else does in the face of the idiot's action.

(Even if the first driver somehow dodges, can the next? What about the guy behind him? It's usually the third or fourth driver back in the stack that starts a multicar pileup I think - aside from distracted drivers that never see the stopped car. The swarm will all slow together when needed, so if the first car makes it there probably won't be an accident.)

Of course, if someone forces a dozen cars to panic stop, the efficiency of the autonomous lane goes way down. Idiots pulling stunts like that are especially the people that need enforcement actions to stop them.

 

[NikeWings]

Not much you can do about idiots like that, except physically separate/block the lane, as some current HOV lanes are.

I was simply pointing out one of many situations on a typical day in the life of Los Angeles HOV, the most highly traveled lanes in the country. It requires a far more complex technological solution be developed, than many understand, and thus the true marvel of our AP as we watch it evolve. Waiting for governments to tender laws or rebuild complimentary infrastructure to achieve the goal, is not an option.

Many HOV and driving studies have been done on CA highways to ensure reality is taken into planning consideration. They are quite interesting to read and encourage you to check them out.

 

[BluestarE3]

The recent fatal accident in which a drunk driver illegally cut into the HOV lane in an attempt to go around traffic and, instead, rear-ended a Model S, grimly illustrates why there's a safety risk in having a rogue driver operating his vehicle recklessly in what is meant to be a controlled traffic lane. I don't see how a well-behaving autonomous car or swarm of cars can avoid being rammed by an idiot suddenly pulling up behind them at high speed and the adjacent lane is congested. Physical barriers will be needed to isolate these lanes from rogue vehicles.

 

[dsvick]

Not much you can do about idiots like that, except physically separate/block the lane, as some current HOV lanes are.

Physicaly separating them would be the best solution, sending them a $200 ticket for each offense would also work. A lot of places have "toll by plate" on toll roads where they take a picture of your license plate and send you a bill. It would be fairly easy to do the same for the automated lane and just send a ticket to the offenders that aren't in an EV. After a couple of those they'll stop, and for those that don't, well they're the once helping to finance the infrastructure for separation.