Traffic-Flow Algorithm Can Reduce Fuel Consumption

thecarchik writes "New projects from German automakers Audi and BMW promise to ease congestion simply by looking at traffic signals and driving style, in an effort to smooth the flow of traffic. Through a test course in Munich, vehicles were able to post phenomenal fuel efficiency gains simply by adjusting the timing of traffic lights depending on traffic volume — to whatever speed provides a so-called 'green wave' of four or more synchronized signals."

Comments

It astounds me

[skelterjohn]

That this isn't done everywhere. With all the red light cameras everywhere (for safety), you'd think they could put a few out there that would make it so I don't spend 3 minutes every morning staring at an empty intersection.

Re:It astounds me

[ktappe]

I'm always wondering if I should go to the city council meeting and ask why they're supporting terrorists with this inane system. The loss in gas mileage is atrocious, and the reason for it is just plain stupidity.

Seems to me the reasons for stupidly-timed lights is threefold:
1) Lowball bids from traffic light installers. To keep their bids low, a simple timer is way cheaper than a smart computer.
2) Politicians who pull strings so their development's side-road gets priority over the main thoroughfare.
3) Citizens like you and me who are too busy to attend council meetings and object.

Re:It astounds me

[Dare nMc]

most of the inductive sensors I have experienced are at the intersection, so you have stopped before they work, thus saving no fuel. They also pick up a single car and only optimize for that one car, and usually only on the lessor used roads, paying no heed to what they are stopping, and for how many. IE the ones I use see a single car (wanting to make a right turn 90% of the time) from a 30 mph lane turning on to a 65 mph highway, and the light will stop a string of a dozen cars going 70. With A very smart camera it would be possible to picking up how many cars, trucks, and where is the next opening. Need dozens of loop sensors to do that.

It would be a huge fuel savings if the lights know for example we have 3 loaded semi-trucks and 5 cars going 70 wait for them to pass and make the slow moving car wait longer. It would also be extremely helpful if we could get info sharing on light timing into something like the google map android phone applications, so that it could tell me to adjust speeds to hit lights, or to create a gap, or turn earlier to avoid a string of bad lights (or join a small group of cars...)

Re:It astounds me

[Anonymous Struct]

I'd add a fourth possible reason, though: I think traffic control may be a little more complicated than we give it credit for. When there's one main road and everybody's on it, it makes sense to try to get long synchronized trains of traffic flowing through green lights. But as soon as you start to get more than one big road, you have to also think about how much traffic you're allowing into different parts of the city at once. If you look at traffic management as a big picture, then giving people green lights doesn't get them off your plate, it just moves them to another part of your grid. If you're stuck at a red light for 30 seconds too long and nobody seems to be going, consider that it may be because 3 miles up the road, that bubble is intended to absorb some traffic from another busy intersection.

Or, as you say, it could just be cheap systems.

Re:It astounds me

[VolciMaster]

On my way home from work, There is one intersection I always have to wait at all the time for about 3 minutes even though its empty. Rights on reds are prohibited, no sensors, just a timer, and it is a 5 way intersection, So even if right hand turns were allowed on red I would need to do right - U turn, Hard Right (Which is actually completely prohibited at that intersection) U turn, Right. It is very odd because every single other traffic light on the road except for that one has sensors, some even having approach sensors changing the light before the cars even reach them.

... You know that if you're biking you can become a pedestrian and cross using the cross walks by pushing your bicycle, and then start biking again, right? :)

that's what I always do when I can't make turns to get where I need to go.

Too Bad It Won't Happen in US

[Avin22]

Here traffic lights are made to be a source of income. They are designed to stop you and increase your chances of running a yellow light so that the cops can pull you over and give you a ticket. Plus, it has the guise of making the roads safer (since people don't have as many green lights, they cannot speed as much), so much of the public is mostly ok with it. Unfortunately, in reality, we're just wasting fuel and making the roads more dangerous (more rear end crashes and angrier drivers).

Re:Too Bad It Won't Happen in US

[tepples]

do you know of any Wal-Marts that don't have a red light?

This one in Fort Wayne, Indiana. It's between a strip mall to the south and some other department stores to the north. The closest traffic signals are two blocks away in each direction.

Uphill battle

[seizurebattlerobot]

Traffic signal timing is nothing new, we've known about it a long time. Unfortunately, there is much money to be made fleecing motorists for traffic violations. As a result, our road systems are tweaked to generate revenue, not expedite traffic. Good luck getting these algorithms used in anywhere but a handful of places without a fight.

Drawbacks of bikes and buses

[tepples]

You want green, invest in buses, trains, bikes, etc.

Being car-free, I know the limitations of bikes and buses. Bikes can't carry a week of groceries for a family of four at a time, and they're uncomfortable in a thunderstorm or the freezing season. Buses in many cities don't run at night, on Sundays, or on national holidays, due to low ridership.

Where the money is

[XiaoMing]

This is somewhat old hat. Companies that depend on urban transportation efficiency for a profit (FedEx and UPS) have long ago implemented systems that recommend routes to drivers. UPS for example uses technology to help reduce/eliminate left turns (usually involve sitting at an intersection idling and waiting, wasting gas and time): http://www.zdnet.com/blog/btl/ups-driving-cost-savings-by-eliminating-left-hand-turns/2190 (2005 article). True it hasn't been done on such a scale or for specifically this exact purpose, but data mining this informational ore vein isn't exactly new.

Off topic, but another slightly more shocking example of just how the drive of money has helped corporations know everything about us: How about being able to predict your marriage and divorce percentage to 90% accuracy? Better yet, how about doing that based on _what you buy_? Visa's got you all covered: http://abcnews.go.com/Business/visa-predicts-divorce/story?id=10320638 ;)

Things like this make me wonder what knowledge about society these companies know about us, and aren't letting ourselves know, simply to help them turn a better profit.

Thinking about drug companies is a scary thought.

Interesting perspective...

[RyanFenton]

The perspective taken for this bit of problem solving is interesting, because it is stepping above the usual street engineering up to city planning - maximizing the number of people able to use shared resources, while minimizing resources used. This is decidedly NOT a perspective that is common in the US, as our cities tend to 'sprawl' at the whim of investors and politicians with 'complicated' priorities rather than anything as idealized as proper engineering to make best use of resources.

Greater use of mass transit to maximize available road where possible, waves of greens with appropriate buffers to keep congestion manageable to even extreme capacities, traffic system that work to inform the driver and minimize late decision making - these are good moves.

I would hope we could use some of these moves to create a road system that would allow for us to approach automated driving systems - where you would decide where you needed to be, and an appropriate vehicle would pick you up within a few minutes, using the minimum amount of fuel for the entire city worth of people using the system, and giving non-automated drivers plenty of road space as they go. Nobody limited in choices - but maximizing efficiency and convenience for everyone.

It probably won't happen here in the US (different priorities, as mentioned), but I hope such a system could be established in my lifetime.

Ryan Fenton

Re:Interesting perspective...

[AK Marc]

This is decidedly NOT a perspective that is common in the US, as our cities tend to 'sprawl' at the whim of investors and politicians with 'complicated' priorities rather than anything as idealized as proper engineering to make best use of resources.

Bullshit. You can design 'sprawl' intelligently, or stupidly. There is no planning in the US, no matter who's doing it or when. NYC is not sprawled, and wasn't well planned. Same with Boston. Dallas is relatively new, and was planned better than many of the older cities and is much more sprawled. It'd better designed than Houston or LA that are more sprawled out it is, though LA managed to deal with the sprawl better than Houston by throwing more and larger highways at it (and it probably doesn't hurt that Californians like to tailgate at high speed, increasing the number of cars in pileups and greatly increasing the throughput of highways).

Look at even bigger cities, like Beijing, with up to 6 or so ring roads and spurs out from the center and it grows more rings as it grows out. They are sprawled, yet planned. So I don't believe that you can't plan sprawl, or, even if not planned, at least modify any existing plan to handle it. What I think happens is that planning means that one politician spends money to make his successor look better. That doesn't win any votes, and if someone from the opposing group gets power, then they will steal the credit, or worse, screw it all up and blame it on your party. So no politicians plan, it's bad politics. And that's a function of the American voters, and not a problem with sprawl.

Re:The green light is "half empty"

[masterwit]

BMW makes cars, which are not "green" by any standard. You want green, invest in buses, trains, bikes, etc. Not more cars.

This is pretty clearly a greenwashing attempt by BMW.

Yes, and I am OK with that...you see they do not make trains, bikes, (or even buses?),...

When a coal company stops mountain-top removal, we acknowledge this and do not disapprove. You do not have to agree with their actions, but even a feeble attempt at smart fuel consumption should be welcomed.

I can see that we need more mass transit "smart solutions", but complaining about some speculation/proposal for improved traffic signals...well cheer up man!

Such a method sounds ripe for deployment on U.S.-style boulevards, where obsolete signals, each running on their own cycle, can bring light traffic to a congested snarl.

That is a true statement, my city fixed its lights, lowered my commute by five minutes+, green washing right into my pocket!

Re:Another reason it's not done in the US...

[tepples]

"Traffic Calming" is how they euphemistically describe things such as lane narrowing and speed humps

Then I've got an even better euphemism: "keeping the crosswalk safer for pedestrians". That's what politicians call it when they want votes.

Re:Poorly designed vehicle detectors

[mister_playboy]

The city I live in has made it illegal to rid bicycles on the sidewalk in the downtown area near the college campus. I am forced to ride in the street in the most heavily trafficked area of town.

I hate it, but it's the law.

Re:Poorly designed vehicle detectors

[jridley]

Roads are for road users. I ride a bike, I obey the laws (ALL of them - way better than ANY car I ever see on the road) and I pay at least as many taxes to pay for the roads as anyone else on the road. If the road isn't designed for my use, it's because the designers screwed up. Bikes were here before cars and they'll be here after cars.

HEADLINE NEWS

[thePowerOfGrayskull]

Through a test course in Munich, vehicles were able to post phenomenal fuel efficiency gains simply by adjusting the timing of traffic lights depending on traffic volume — to whatever speed provides a so-called 'green wave' of four or more synchronized signals."

This just in! Stopping and idling at each of four consecutive lights uses much more gas than driving straight through them without stopping!

Re:Another reason it's not done in the US...

[thePowerOfGrayskull]

The leading cause of death for people between 15 and 35 in the US is automobiles.

You sound like a big part of the problem.

The leading cause of death is not automobiles, just like the leading cause of homicides is not guns.

Every single automobile death is caused by either someone not paying sufficient attention (which includes driving faster than you can plan ahead for any given set of road conditions), or mechanical failure (usually coupled with some degree of someone not paying attention). The idiot who pulled out "right in front of me" today might have caused an accident -- he certainly wasn't paying attention, but I was. I saw the potential for him to make a stupid choice about four seconds ahead of time (as soon as he approached the intersection*) and by the time he actually did it two seconds later, I had already dealt with the situation.**

The point is this: most accidents require two parties not paying attention: the one who is making the active mistake and does something stupid like pulling into traffic, or driving faster than he can react; and the one who should have seen the possibility for the event to occur, and reacted to avoid it. Obviously this doesn't apply in all cases - but I suspect it does in most. (Heck - even getting t-boned when crossing a protected intersection ... why did you assume you didn't have to look first, just because your light was green?)

* what is with those people who will pull up to an intersection, sit and STARE at oncoming traffic for several seconds, THEN pull into it, causing much swerving and slamming on of brakes? Are they just spaced out while they watch a few dozen tons of metal bearing down on them, or perhaps they know they *should* look left -- so they do it for form's sake though they've already made up their minds to go?
** no, "dealt with" does not mean slamming on the brake and potentially causing more incidents behind me. With so many more controls than the horizontally long pedal in the middle (or left), WHY is that so often a person's first reaction?

As far as GP goes - I feel his frustration, and it's nothing to do with wanting to speed. Driving in a residential area I only drive as fast as I can comfortably react to the unexpected - often given the potential for kids and animals coming onto the scene from hidden places, it's at or only slightly over the speed limit. But when I hit six lights in a row, and the timing of the lights causes yet more delays and backups because of volume... i can't help but get frustrated at what a waste of time and gas it is. Those lights *could* be sequenced together, but they're simply not -- usually in order to either raise ticket revenue for the city, or because some politician who doesn't know anything about traffic management decides something must be "done" about how unsafe it is in the city.

Re:Poorly designed vehicle detectors

[tsm_sf]

I'm not going to argue with your point of view (since you're more or less correct), but I would like to point out that offensive (as opposed to defensive) riding is unfortunately essential in a large urban environment. Getting out in front of traffic at a stop light, riding two abreast, moving from sidewalk to road and back again, and splitting lanes are all part of not being hit. For every asshole bunny-hopping the curb or cube-gleaming your fender, there's another who's totally oblivious to the fact that they're behind two tons of metal.

It's a lose-lose scenario, yet for some reason blame is passed between various forms of commuters instead of placed on urban planners where it so rightfully belongs. We need one lane for transit/commerce/utility, one lane for personal motorized transport, one for muscle powered, and one for pedestrians. Crying that it's difficult is begging the point.

((Critical Mass doesn't fit into this picture. It's a protest, and you know full well when and where it will occur. Getting your side mirror bashed or your windshield krypto'd is your own damn fault))

Re:Greenwashing

[mirix]

Cars aren't going away any time soon. So we can:

A. Do nothing.

B. Fix the traffic lights for minimal cost and offer some improvement on things.

But I guess since B doesn't remove cars entirely, we should do nothing right? That's pretty fucked logic you've got there. If doing this saves only 10% on urban fuel consumption, it will have the same effect as 1 out of 10 people stopping driving entirely. Seems like a net positive to me, and a lot more feasible than hoping 10% of people to give up their cars and start walking everywhere.

My route to work is horrible. I hit nearly every light, every day, even when coming home in the middle of the night. I'd like to send the city a bill for 20%+ of my gas, and half the cost of replacing brakes & clutch when the time comes, as this could have been easily saved by fixing the fucking lights. The rage induced by hitting every light probably knocked a few years off my life too.

Re:Poorly designed vehicle detectors

[driptray]

The argument that cyclists don't pay their way is absurd - the reality is the opposite; public funding of roads is a massive subsidy from those who don't drive a car to those who do. As a cyclist I'd love a system of user-pays for roads.