Poll & Prediction: Autonomous Car Equipment at 5k by 2019

[DrWho]

I think the inefficiency I see on a daily basis is that people won't yield, so as they try to move over a lane, they have to slow down because no one will let them in, and it has the effect of slowing down that entire lane for some distance behind.

 

A couple of thoughts....

 

What formal training is already provided in regards to merging? In Illinois it doesn't amount to much more than a lecture about using a turn signal and waiting for someone to make room. That's certainly not what my parents taught me, and certainly not the unspoken rule of the road in Chicago. I would suggest that education about merging is severely lacking - if for no other reason than everyone has their own opinions about it. I keep harping on the education thing because Germany doesn't have the merging problem - which means the problem isn't inherent to human operators.

 

As far as merging goes, how should it look in an ideal world? Again, you're going to find people with different opinions. Every method, however, boils down to the fact that someone needs to slow down. The emotional response encountered when you see someone "not doing it your way" is not a quantifiable metric that can be discussed. I've had several discussions with people about when to merge when a lane is closing and I've personally been in both camps and felt the emotions both ways. I don't want to argue every single specific scenario because it's missing my point, but I want to point out that these things can be quantified. Since merging was brought up, let's use an example with a 3 lane road going down to 2 lanes with equal traffic density and flow in every lane. Some would argue that you should merge at the last possible second (Option 1), some would argue you should merge as soon as you known about the lane closure (Option 2). I think most would argue that once they decided it was time to merge, that the car behind them shouldn't then pass them and merge ahead of them (Option 3 which I'm calling the emotional argument). Let's start with option 1, which causes a 1/2 mile backup. If the lane closure notice was made 1 mile in advance, then the backup caused from Option 2 would actually be 3/4 mile. This means that the approach from Option 2 would need to have a faster flow rate to offset that backup difference. Will it really be 50% faster than the Option 1 flow rate? The interesting thing here is the output flow rate after the lane reduction. If there is a higher flow rate going in than what can be supported by the lane reduction, then you will always have a backup regardless of the merge technique. In this case, Option 1 is always better because it minimizes the area of the impact (other exits are no longer being blocked). If you're following the logic here, then you want the merge to be as late as possible to where the input flow rate matches the output flow rate. That necessarily requires that cars merge throughout a spectrum between the notice and the actual lane closure. The thing that is fascinating about this is that there is a range of "aggression" on the road....the more aggressive drivers merge later, and the more passive drivers merge earlier. This is a natural way of creating that spectrum without formal regulation. The emotional response isn't as nice though because if you're an average aggression driver, then you feel like half of the people are driving too aggressive and half are driving too passive. Unless you're the most aggressive driver on the road, then you're always going to find yourself in the Option 3 camp....because you're driving as fast as you think is safe. And this gets back to my education scenario....how can everyone on the road have a different perspective of what is safe? (Beyond differences in the capabilities of different cars). The point is that dude you're complaining about needs to be driving that way to optimize the flow.

 

Anyways, I've rambled way too much. What I wanted to emphasize is that you can start ascribing numbers to quantify various approaches, and then offer models that should provide improvements. The reality is that most traffic backup scenarios are caused by flow rate limitations....not merging technique. Get off the "cutting in line" perspective and just look at the distribution of cars and their speeds and the ripple effects being caused. Rubber-necking is by far the bigger merging issue - which is why that has been the focus of transportation experts. The fact that the rubber-necking scenario isn't the first one brought up when trying to defend AVs shows me that someone hasn't been looking at the data - what better example of a human reaction causing real traffic problems? Anyways, the thing about these models is they don't care if the drivers are human, dogs, robots, whatever....there is finite road with gaps and speeds limited by the nature of the road itself and the cars driving on it. If there were a way to improve flow by even 20%, then those methods would be implemented long before adding more lanes....

[Gilbert]

Can someone explain to me how a 5 mile traffic jam suddenly becomes 1 mile ("5x more capacity") just by switching to AVs?

 

 

According to Dave it's suppose to be simple, but I can't figure it out either.

[Mallette]

According to Dave it's suppose to be simple, but I can't figure it out either.

 

As you say, it's simple.  It won't happen in the first place.  Only a car driven by an idiot would go to where there is a traffic jam when there are "x" numbers of ways from "a" to "b."

 

Dave

[Gilbert]

 

According to Dave it's suppose to be simple, but I can't figure it out either.

 

As you say, it's simple.  It won't happen in the first place.  Only a car driven by an idiot would go to where there is a traffic jam when there are "x" numbers of ways from "a" to "b."

 

Dave

 

 

 

But that would require constant, real time monitoring of traffic and road construction (to name just 2 variables) for all the possible combinations of "X".... which can be done, it'll only cost a small initial investment somewhere in the vigintillions, per city.

 

Which is why totally autonomous ain't gonna happen in our life time.

Edited April 14, 2015 by Gilbert

[Mallette]

THINK, Gil...no, it doesn't.  The system to provide real time traffic data is already in place and many of us have it on our phones and/or navigation systems.  The difference is what the AVs already do...talk to each other and pass on data.

 

It really is simple, and nothing not already with us is necessary. 

 

Sheesh, another wreck at Scarborough this morning.  Like deja vu all over again.  I really thought the traffic info sign had somehow not been updated since yesterday morning but when I approached my exit traffic was coming to a halt ahead of me, just like yesterday at the same time. 

 

My car would have known that in advance as well as which exit would be best to get while the getting remained good.  This is NOT rocket or future science...it's right now. 

 

Dave

Edited April 14, 2015 by Mallette

[Travis In Austin]

An integrated system that coordinates AVs with traffic buildup is certainly much further down the line. This can happen in major cities but it will come with the cost of additional time. If freeways are congested it will put you on surface streets, alternate routes, etc.

The model will probably involve tolls and fees for Peak times. A toll lane in Austin begins next year that works that way. The toll is variable, from $1 to $5, depending on the congestion level for about 10 miles of road that goes directly into downtown.

An integrated AV system will probably give you alternateatives, freeways and less time would cost X, surface roads and longer routes would be Y. Less if you travel between 10 and 2.

UT is working on intersection and traffic control design for use with AVs. They have been asked to work up a model that gives intersection priority to emergency vehicles and those vehicles that bid for and receive priority. This would be delivery trucks, FedEx etc. As you approach an intersection and a company truck approaches perpendicular, you stop and the truck goes through. The truck that has bid and paid gets the "green" light.

There will no longer be income from traffic tickets, municipalities will have to make up that income in other ways such as this.

I think the big question is regardless of the technology, full AV, or just partial driver assistance, at what point will we see a 10 percent improvement in commute times? What percentage of vehicles in urban areas need to have the tech to see any real commute time improvement?

I think the commute times will continue to worsen as already projected and then slowly improve down the road. I think ten years is optomistic on when improvement begins, but who knows.

As always, it depends on how much of a legislative priority it is, the ever present urban vs. rural competition for funds, and how this ends up being legislated on a local,state and federal level.

As always, time will tell.

Travis

[Mallette]

Well, I still don't see anything needed not already in place.  However, rather than debate that, I'll fully agree that it's going to be interesting to see at what point the mix really starts paying dividends one can experience.  The process is already started with a tiny number of vehicles that can take over in traffic jams and keep you in a lane.  Personally, while it may not get me there any faster I'd sure be a LOT happier just not having to do the inch, inch, thing for half an hour or so. 

 

Those systems alone are going to sell a LOT of new cars.

 

Dave

[Travis In Austin]

I agree Dave, the hands free alternative in rush hour is well worth it, just in terms of stress reduction alone.

There are major changes that need to be taken to have a fully integrate connected AVs. UT and A&M are doing studies and designing systems. This will control the speeds of AVs on surface streets, monitor flow in real time (which is already being done right now with cell phone pings on an anonymous basis), will factor in anticipated loads based on destinations people have entered into their av, etc. Eventually, AVs will be required in congested urban areas, and eventually there will be no traffic lights, except to tell pedestrians when it ok to cross the street.

That is if, and when, cities want the technology to be the most efficient.

[Gilbert]

Sorry to hear about the traffic, and I'm not looking forward to having to drive all the way back to traffic hell in Houston, just so I can look pretty and entertain the municipal court judges questions regarding why I refused to pay the toll. 

 

I don't  think "Failure To Pay Toll" is even a remotely accurate description of my violation, because if there had existed a toll booth I would have gladly paid the fee. And if the Toll Authority (or whatever you call them) had not designed their toll lanes as traps, I would have exited the toll lane.

 

To top it all off, some shyster Law Firms are sending me solicitation notices regarding the toll ticket. I can't begin to understand how they find this shyt out, and assume they have someone living in the dumpster over at the circuit courts. The best price is $50 bucks, for $50 bucks they will represent me. I actually called one of them, the guy was flapping his lips 100mph, but I cut to the point, and just asked if I I paid him the money, would I still have to go show up for the hearing..... he said "yes", and said thank you very much, and hung-up the phone.

 

 

Concerning the "Systems already in place", sorry, I don't see this being even remotely as simple as you are assuming. One hick-up in autonomousville, could negatively impact 1000's upon 1000's of traveling lemmings, they could end-up snfu'ed for hours because some sensor went snafu,.... no, I do not believe the intermittent systems scattered about and currently in place are enough. Just the thought of one of those hick-ups causes a chill down my spine. You Houston boy's are already fk'ed......  I hate that place.

 

Your city is like New Orleans, you're in a huge traffic hole, and the levy was compromised decades ago. It's time to pack up your shyt and get the hell out of there before it's too late.

Edited April 14, 2015 by Gilbert

[Mallette]

Well, you see Gil, if you had an AV you wouldn't have to show up...it would. 

 

Anyway, I think the cities looking at systems may not really understand the technology.  These vehicles are being designed to talk to each other and pass on the information.  The code would be written so that, of course, priority one is to the vehicles within a short distance, two is to conditions ahead, and three are conditions all the way to the destination.  All of these would come primarily from the other vehicles on the road passing them vehicle to vehicle. 

 

Nothing new needed, all within existing AV needs.   

 

Dave

[Travis In Austin]

Which toll road is that down in Houston that doesn't do toll by mail? I didn't even know that existed.

All of the toll roads up this way are mail or TxTag, you save a little if you do TxTag.

What a scam, that same thing happened to me in Denver, no pay lanes. I sent a check for the amount of the toll, and put payment in full, they cashed it.

Sent that to the contractor that they use to collect and I never heard from them again.

Edited April 14, 2015 by dwilawyer

[oscarsear]

Still churning this topic I see.  Read a study reporting that in upwards of 15% of people who ride in autonomous vehicles will get motion sickness.  Hope they come with barf bags.

 

Carry on.

[Mallette]

What a scam, that same thing happened to me in Denver, no pay lanes. I sent a check for the amount of the toll, and put payment in full, they cashed it.

 

Actually, the entire SE quadrant of Beltway 8 is tag only, as is Westpark and several others.  I think it sucks too.  I've a tag, and those segments are nice as many fewer use them...but it makes no sense.

 

Dave

[Mallette]

Still churning this topic I see. Read a study reporting that in upwards of 15% of people who ride in autonomous vehicles will get motion sickness. Hope they come with barf bags

 

Happens to a lot on their first flight, and happened to many from what I've read on their first train rides in the 19th century and first automobile rides.  They'll get over it.

 

Dave

[Gilbert]

Which toll road is that down in Houston that doesn't do toll by mail? I didn't even know that existed.

 

 

All of Houston's Toll Roads are like that. I mentioned this very thing to the young lady who assisted me at the EZ Tag store. She said, "we don't have that here". Freaking Dallas at least has toll booths where you can pay, Houston gives you a dic in the arse, with the cherry on top being a ticket requiring a mandatory court appearance. It's no wonder Dave wants an autonomous car. Driving there sucks rocks, big time. I'd hate to think what it's like over by NASA during rush hour. OMFG, IH45 btwn Galveston and Houston is suppose to be a biach, and I've always left Galveston taking the back roads to Schulemburg, so I'd pop out approx. 1.5 hrs. west of Houston, and well away from traffic hell.

 

Anyway, I was told that the EZ Tag works on all Texas Toll roads, including my favorite, State Hwy 130 between Seguin and Austin. They say it goes all the way to Dallas, but I've never taken it that far.

Edited April 14, 2015 by Gilbert

[oldtimer]

At least in Dallas if you don't have a tag they take a picture of your plate as you pass by and eventually mail you a bill.  It sounds incredibly wrong to me to have a system that issues violations for failure to pay without giving you a chance to pay first.  Give'em hell, Gil.

[DrWho]

According to Dave it's suppose to be simple, but I can't figure it out either.

As you say, it's simple. It won't happen in the first place. Only a car driven by an idiot would go to where there is a traffic jam when there are "x" numbers of ways from "a" to "b."

Dave

So quantify it Dave, pick a traffic jam in Houston and then show us the alternative routes.

The fact is those alternative routes don't yield 5x capacity. If they did, navigation systems would already be directing people along those paths. The flow rate of the side streets is simply lower. It's that way by intentional design, and that approach will be around for many many more years.

I'm surprised you're not arguing that people shouldn't be leaving their house in the first place. All this newfangled technology....why not work from home?

I don't think we desire the same human existence...

[DrWho]

Still churning this topic I see. Read a study reporting that in upwards of 15% of people who ride in autonomous vehicles will get motion sickness. Hope they come with barf bags

Happens to a lot on their first flight, and happened to many from what I've read on their first train rides in the 19th century and first automobile rides. They'll get over it.

Dave

Haha, don't get in the way of Dave's vision. Der Fuhrer hat gesprochen!

I wonder how many of those people simply don't take trains, busses, and planes? I know two people close to me that get sick anytime someone else is driving. Or maybe it's just when I'm driving?

[oldtimer]

Drive them out to the woods and leave them there.  Round up your friends with peanut allergies for the same while you are at it.

[Mallette]

So quantify it Dave, pick a traffic jam in Houston and then show us the alternative routes.

 

Wreck at Preston Road/La Porte Freeway this morning.  I had two viable routes to get around it.  Without the data from other vehicles to help I had to guess.  Luckily I saw the info sign and knew I'd be able to see the backup well before getting there.  But I still had to choose either the Jesse Jones bridge over the channel or the service road.  Made the right choice and my delay was minimal but would have preferred not to have to guess.  There are always alternatives except at bridges and a few other places.  Even then, when traffic is down to a lane or two the complete "me first" attitude of most drivers here so slows and creates such stress...and often secondary accidents...as to cut the potential speed of getting through in half or worse.

 

Dave