backtop

Print E-mail del.icio.us 41 comment(s) - last by heulenwolf.. on Nov 17 at 10:41 AM

Road Train Diagram  (Source: Wired)
Road trains would require vehicles with technology already available

A research project is underway in the European Union that looks to devise a way that drivers can carpool without having to ride in one vehicle and still get the benefits of fuel savings and fewer accidents. The project is evaluating the use of so-called road trains.

The road train would be controlled by a lead vehicle, likely a professionally driven vehicle like a bus. Wired reports that the project is going by the acronym SARTRE and that no infrastructure investment will be required to enable the road trains.

“By developing and implementing the technology at a vehicle level, SARTRE aims to realize the potentially very significant safety and environmental benefits of road trains without the need to invest in changes to road infrastructure,” project coordinator Tom Robinson said in a statement.

The fuel savings are estimated to be in the area of 20% and drivers in the road train would be able to take their hands of the wheel and sleep or eat while they are driven to their destination in their own vehicle.

“This type of autonomous driving actually doesn’t require any hocus-pocus technology, and no investment in infrastructure,” said Erik Coelingh of Volvo, one of the principal companies participating in the project. “Instead, the emphasis is on development and on adapting technology that is already in existence.”

BBC News reports that the project is being financed by the government as a way to reduce accidents and ease congestion on the highways. Robinson says that the idea is to use off the shelf components to enable vehicles to participate in the road trains. The vehicle setups would require a GPS navigation device and a system to control the steering of the vehicle. Vehicles would be able to enter and leave the road train at any time. The first test vehicles are expected to debut in 2011.

Robinson said, "Each of the vehicles will have their own control and software monitoring system. There may well be a platoon sensor envelope that collates information and presents it to the lead vehicle so it can understand what is happening around all the vehicles."


Comments     Threshold


Great Idea
By Connoisseur on 11/11/2009 12:59:56 PM , Rating: 2
This sounds like a great idea. A system that leverages the advantages of drafting while still providing owners individual control to go to specific destinations. I see a couple of hurdles for this though:
1) They'll need a system to control ingress and egress especially if there's other, manually controlled traffic in the various lanes.
2) I'm assuming that, when in this train, the only system controlling the organization of the vehicles is a computerized GPS nav system. That seems to be a lot of faith to put into commercial system that's only accurate to a few feet. A safer option would be some kind of physical hitch/hook system so that the cars can be as close as possible and still move/stop with the lead vehicle.

All in all though, seems the best of both worlds (individual vs. mass transit).




RE: Great Idea
By HotFoot on 11/11/2009 1:07:34 PM , Rating: 2
They could use ultrasonic sensors for automatic relative positioning of your car versus the vehicle in front of you, but keeping the whole thing on the road through sharp turns and that sort of thing, I have to think, would take some kind of awareness of the geometries of the road as well.


RE: Great Idea
By Smartless on 11/11/2009 2:21:35 PM , Rating: 2
Yes with today's technologies and UI's in dashboards this is much more plausible. The theories and ideas have been tossed around by traffic engineers before to increase capacity on freeways (so sharp turns are less of a problem). Since today more cars are computer controlled, decreasing headway between cars and having trains like these would theoretically allow for a much more efficient and safer road while also having the freedom and privacy of your vehicle. I for one just want a digital finger to appear in the windshield of someone who cut me off.


RE: Great Idea
By grath on 11/11/2009 11:06:39 PM , Rating: 2
quote:
some kind of awareness of the geometries of the road


Machine vision to detect road surface markings has been around quite a while, I remember seeing (seemingly) working prototypes demonstrated on Discovery Channel like 10 years ago.

Basically it used used a forward looking camera with the contrast set very high so that an image is generated of only the borders between objects and colors, then it filters out any borders that are not near-parallel to each other, and the result is a fairly clean image of just the road surface markings. The computer constantly plots a future target position between the lane dividers and adjusts steering toward it to keep you in the lane.

That assumes you have an unobscured forward view of several seconds at least. For this vehicle train, the lead vehicle would have the best view and could relay it to the rest of the train, and the individual passenger cars could have side mounted cameras to get a better view around the car in front of them, and provide some redundancy for keeping the train in its lane.


RE: Great Idea
By Keeir on 11/11/2009 1:37:01 PM , Rating: 2
quote:
All in all though, seems the best of both worlds (individual vs. mass transit).


Funny. I was thinking its the worst of both worlds.

Additional Cost through having a "professional" driving out driving the center communication equipment around. Not only would the center cost alot of money, but there would likely be significant maintainence and on-going labor costs. The fuel costs of this may be so significant as to render drafting from even 5+ cars as not efficient.

Additional Cost through GPS/Software systems installed on people's car.

Additional Risk because having a *unknown* length "train" in front of you or on the side of you will cause significant problems in the surrounding drivers decision process.

Additional Risk through having a driver be -not in control- of his own car for an *unknown* length of time. I am sure several will just daze out and problems will result (no matter how loud and persistant we make the warning signals).

Maybe because its the UK, there would be sufficient density of these to make a difference, but in the US I would see having a great deal of diffuculty even finding a train going close to where you want to go...

All in all seems like a fail

I still have to do the worst parts of having a car: I need to maintain my car, put miles on it, exponse myself to liability, AND still pay parking and gas. I still will have significant limitations of a public mass transit system: I will get to my destinations slightly slower (at best), may be forced to take a non-optimal route (if I want to be part of the train), and worst of all, still paying a premium for the mass transit whether I use it or not...


RE: Great Idea
By TheDoc9 on 11/11/2009 3:02:44 PM , Rating: 2
Yeah, it's doomed. I only see this being a positive for drunk drivers. That's assuming the drunk can find the train and navigate into and out of it. It might also be big during the holidays when people are on long trips.

I would also have a concern about the professional drivers. From the graphic one would assume you'll be following a bus driver. Good luck!


RE: Great Idea
By Jeffk464 on 11/11/2009 4:58:15 PM , Rating: 2
I was thinking about the same thing with drinking and driving. The only problem is that the driver has to drive to a certain point to "hook up" to the train, then drive again to get to their final destination.


RE: Great Idea
By johnsonx on 11/11/2009 9:15:55 PM , Rating: 2
yes, but anyone who's driven when they maybe shouldn't have can tell you is easy to maintain your attention level and control of the the vehicle for a few miles... it's the extended drive where being tired and the buzz cathes up with you.

I'm not saying that makes it ok, simply that what you stated isn't really the problem. The bigger problem might be that this could encourage drunk driving... "if I can just make it to the train, then I can sleep this off until I get home".


RE: Great Idea
By Connoisseur on 11/11/2009 3:37:11 PM , Rating: 2
I dunno. I recall mythbusters doing an episode about drafting behind a big rig. If they can design a system which will keep all the vehicles within about 5-10 feet of each other, they may encounter significant improvements in fuel economy (~20-30%). As long as the cost per "bus" (including driver salary, maintenance, fuel etc.) is less than the fuel economy savings of say 5-10 cars, it could be a feasible venture. As far as having to stick to a route, I would say this this "road train" should be a strictly optional exercise. I can foresee this being especially useful during long (interstate or international) trips when you just want to sit back and eat a snack or take a nap. I'd say the vast majority of the people driving experience some form of driving fatigue on >3 hour trips.

Of course... this is now delving into the realm of fully automated vehicles.


RE: Great Idea
By Jeffk464 on 11/11/2009 5:01:05 PM , Rating: 3
If it gets good enough to have a huge decrease in accidents you could gain further fuel improvements. If you could get accidents to say aviation levels then you could dump a lot of the safety features in cars. This would reduce the weight of the car and reduce the size of the engine needed so should give a big boost.


RE: Great Idea
By Keeir on 11/11/2009 7:22:11 PM , Rating: 3
quote:
I'd say the vast majority of the people driving experience some form of driving fatigue on >3 hour trips.


Which is the rare condition. 95% of most driving is less than 1 hour at a time, let alone 3 hours plus.

quote:
As far as having to stick to a route, I would say this this "road train" should be a strictly optional exercise


Which is why its doomed. Would you pay an 1,000 dollars (or so) to upgrade you car and pay the other fees if you can't be reasonably certain that on your 2-3 times a year when you would want to use this service for long distance travel that it is even availible? Would you be willing to wait 15-30 minutes on a 2 hour journey to hook up to the "train".

And before you get any ideas about every truck and bus using the system, consider the liability involved. Most private companies will want no part of this without -significant- payback.

I mean, I think the system is cool and all... I just have trouble imaging it being practical in any US cities.


RE: Great Idea
By Alexstarfire on 11/12/2009 2:10:42 AM , Rating: 2
The point was that you wouldn't have to upgrade most cars. And I would certainly wait 15-20 minutes on a 2 hour trip. Wait 15-20 minutes to get 2 hours of free time, why wouldn't you?

I don't see this as really being feasible though. If you can get this system running then why not just have fully automated cars. They'd use the exact same principles, on the highway anyway. Which is where I'd assume this would be running since if you had stoplights you could split up the train or have a TON of people running red lights.


RE: Great Idea
By Mint on 11/11/2009 4:09:34 PM , Rating: 2
Fuel economy is just a side benefit. The real reason this will catch on is for convenience. It's the first step towards fully automated driving, and that's why people will pay a fee for it. The lead bus could be part of an already existing public transportation express route.

Risk will be reduced, not increased. No lane changes, no excessive braking, no idiot drivers, etc.


RE: Great Idea
By Keeir on 11/11/2009 7:12:28 PM , Rating: 1
quote:
The real reason this will catch on is for convenience.


Like mass transit is so convient? I can only seen this working where mass transit currently does... which begs the question, why not use mass transit?

I just can't envision a real-life US system which would work in most cities.

BTW for people thinking about Safety. I have never seen safety statistics broken down to the Bus level on a national basis, but the statistics say Larger Vehicles get into -more- accidents on a miles traveled basis than Small Vehicles. Of course, Larger Vehicles are have low fatalies/injuries per mile and crash. I think an "extra-long" bus will encounter more crashes per mile than a passenger car. Since the people involved will still be in lightwieght passenger cars, fatalities/injuries among Car Train Systems will likely be higher per mile than normal driving.


RE: Great Idea
By JediJeb on 11/11/2009 6:46:16 PM , Rating: 2
I dont think the control center would be in the lead vehicle, but consist of just a beacon to activate everyone else. The steering and following distance controls would be in each individual vehicle. Also "professional driver" could just include a truck driver. Imagine on a US Interstate if they just pay truck drivers a little extra per mile to lead people from one city to the next, the big truck causes the initial vortex for drafting, all the cars link in the proper formation to take advantage of it and just follow each other. If one leaves the train the rest accelerate and take up the slack and resume drafting. The lead truck would transmit a braking signal if there is an emergency ahead and the rest use auto braking and distance controls to avoid collision. Not totally simple, yet with the computing power today not unthinkable either.


RE: Great Idea
By Strunf on 11/11/2009 7:28:50 PM , Rating: 2
The problem is when the lead vehicle makes an emergency break, all the cars on the train will crash and since the distances are soo small it would be as not breaking at all.

Thinking about it I doubt this kind of thing would be allowed, there's a minimal distance you have to be from the car in front of you depending on your speed (so you can break) and if you are closer than that you can get fined, everyone on this train would be violating the "law".

Also what would happen if you try to overtake one of these trains and before you manage to overtake it a car comes on the opposite direction then you would be in a dangerous position.


RE: Great Idea
By Alexstarfire on 11/12/2009 2:16:11 AM , Rating: 2
You forget this would be simultaneous braking, so the distance between identical cars wouldn't matter. Though I foresee problems with the varying sizes of cars in the train. A sedan stops faster than an SUV. So if an SUV is behind a sedan it could cause problems. Though with the distance being so small between the cars it's be less of a collision and more of a bump and then the sedan uses it's braking power to slow the SUV down as well. The impact force between cars is caused by the relative speed difference between them. So if one car is going 70 MPH and another is going 71 MPH is almost like hitting a parked car at 1 MPH, assuming they are traveling in the same direction.


RE: Great Idea
By JediJeb on 11/12/2009 9:53:28 AM , Rating: 2
Exactly! The 2-3 second safety distance requirement between vehicles is to allow for human reaction time not actual vehicle stopping distance. If the lead vehicle is a large truck or bus, it will not be able to stop as quickly as a regular car or pickup so they should never collide with it if they receive the braking signal the instant the lead vehicle brakes. G-force sensors can determine how severe the braking is and react accordingly.

As for the different vehicles not stopping in time, well that should be taken care of by each vehicles onboard controls. If each vehicle is programmed with its braking characteristics then it would know how far to follow the vehicle in front of it. Place a transmitter on the rear of each vehicle that transmits the same information to the vehicle behind it and you get another level of accuracy to add to the proper following distance.

The real problem comes when the human driver does not trust the automated system and tries to tweak it by increasing of decreasing the following distances or braking harder when it doesn't need it. Vehicles already are being made with front and rear proximity sensors, alarms, emergency braking systems, and lane positioning systems ( Mercedes is one I believe with these, maybe more).

Plus you don't need full auto drive if the lead vehicle has a driver. You also won't need drivers on assigned routes as you could have many vehicles with professional drivers ( trucks and busses as stated above) Who can activate a beacon to follow when traveling through a city or between cities, then the individual vehicles use GPS where you set the exit you need and it will drop you out of the train at the appropriate spot, signaling the following vehicle you are leaving which will then take up the gap and continue on. If the lead vehicle needs to exit, it send a signal to the following vehicles to disperse the train or hand over lead to the first following vehicle if it is qualified. There are many ways it can be made redundant and useful without requiring a "train driver" following a circuit.


RE: Great Idea
By guacamojo on 11/12/2009 11:29:26 AM , Rating: 2
quote:
Exactly! The 2-3 second safety distance requirement between vehicles is to allow for human reaction time not actual vehicle stopping distance. If the lead vehicle is a large truck or bus, it will not be able to stop as quickly as a regular car or pickup so they should never collide with it if they receive the braking signal the instant the lead vehicle brakes. G-force sensors can determine how severe the braking is and react accordingly.


In grad school (~1997) I saw a presentation by a research group which had outfitted a pretty basic car with a (then-new) active cruise control system. The test had a lead car which was put through a sequence of acceleration and braking (including hard braking) maneuvers on a closed highway, with speeds ranging from ~5 mph to ~70 mph.

Through the test, the follow vehicle maintained a following distance of 1 meter, with a maximum error of 10 cm (10%). That's with no radio communication between vehicles, just using range detection. Radio links between lead and follow cars would have improved that performance, because a distance error wouldn't be required to effect a correction.

If the lead vehicle is a bus or other large vehicle with limited acceleration and braking, then normal passenger cars wouldn't have any problem maintaining following distance.

I'd think there would be some interesting control engineering problems with avoiding following-distance oscillations in a long train of cars, but the radio links would probably help with that, too.


RE: Great Idea
By Keeir on 11/11/2009 7:29:23 PM , Rating: 2
quote:
Not totally simple, yet with the computing power today not unthinkable either.


http://en.wikipedia.org/wiki/DARPA_Grand_Challenge...

When they make a million dollar car able to complete these test at normal driving speeds (60 MPH for Grand Challenge, 30 for Urban), maybe I can have a little more faith that relatively cheap systems and software have a chance on the road with humans, weather, and other road conditions.


RE: Great Idea
By log on 11/12/2009 6:18:51 AM , Rating: 2
My interpretation was that the professional driver would be a bus or long vehicle driver, driving existing commercial vehicles. It would not be about having dedicated drivers and vehicles to head these road trains.


RE: Great Idea
By Murloc on 11/12/2009 10:10:55 AM , Rating: 2
well they could use the new european localization system which will be better.


Manual "Speed Train"
By Yawgm0th on 11/11/2009 4:52:10 PM , Rating: 1
Get behind a semi or bus every commute and follow until your exit. I do this almost every day and get a solid 5% boost in MPG. Getting dangerously close gets it even higher (assuming you don't have to break constantly), but I think I'll pay the extra dollar or so a month in gas to not rear-end another vehicle.

The method being considered is an application of slipstreaming, a well known hypermiling technique. I just don't see any sort of automated application of it as practical or safe. If the EU manages to get this implemented, great -- but I wouldn't hold my breath.




RE: Manual "Speed Train"
By Jeffk464 on 11/11/2009 5:03:52 PM , Rating: 2
The vast majority of accidents are driver error. If you can take the driver out of the equation, because we all know they aren't going to improve, then you can reduce accidents.


RE: Manual "Speed Train"
By mdogs444 on 11/11/2009 6:45:41 PM , Rating: 2
Sure, but would you trust riding in an airplane every day with no pilot? Would you REALLY trust the car computer with your life?

I wouldn't, and I'm sure I'm not alone.


RE: Manual "Speed Train"
By Alexstarfire on 11/12/2009 2:22:53 AM , Rating: 3
Planes are quite capable of running on autopilot from takeoff to landing from what I've read. I think it's specifically after takeoff, but I don't quite remember. I wouldn't be surprised if most of the time a plane was in autopilot mode with the captain just sitting there making sure nothing out of the ordinary happens.


RE: Manual "Speed Train"
By stromgald30 on 11/12/2009 5:29:16 PM , Rating: 2
Cars are quite different from planes in terms of complexity of automating. Aircraft have more control methods (ailerons, spoilers, elevators, rudder), but there's a lot more room for error or failure (i.e. stalling and crashing).

With cars, vehicles are packed much more closely together, and the distance from obstacles are tiny compared to the buffer that aircraft have.


By sciwizam on 11/11/2009 12:37:21 PM , Rating: 3
Gah!! Crash!




By Fred242 on 11/12/2009 3:22:06 AM , Rating: 1
If you have a train of say ten vehicles driving close enough to give a significant aerodynamic benefit so they are all effectively enclosed in the same bubble of moving air, can you imagine how poor the air entering car 10 will be? You will be breathing almost pure exhaust fumes. The cars themselves will suffer almost as much both in terms of air quality entering the engines and in terms of waste heat and lack of airflow to the radiators for cooling. I give the train about 2 miles before you either asphyxiate the last driver, asphyxiate his engine or terminally boil his cooling system.


By tmouse on 11/13/2009 9:49:54 AM , Rating: 3
On what logic do you base that? If a crash occurs and effects the lead vehicle all control to the entire convey is lost immediately. All of those people napping or eating have at most a couple of seconds or less to respond. I suppose if communications are lost a break signal could be sent to all cars but it would have to be a hard brake and then you have multiple whiplash victims. Practically every accident will involve multiple vehicles. You also will probably have increased accidents caused by people trying to pass the train hitting others trying to leave. I know you can have 3 lanes (passing, train and exit) but at least here in the US that will not work.


Sweet!
By Motoman on 11/11/2009 1:01:28 PM , Rating: 3
It's just like in that one Andromeda episode when they slaved all the ships in the fleet to the Andromeda to navigate that asteroid belt, or whatever.

Pew pew! No QQ!




RE: Sweet!
By Helbore on 11/11/2009 2:07:02 PM , Rating: 2
Well if it's anything like Andromeda, that proves it's a BAD idea!

I don't want Kevin Sorbo driving my car, thank you very much. ;)


RE: Sweet!
By Motoman on 11/11/2009 3:37:23 PM , Rating: 2
Maybe not Kevin, but Lexa? That would be fine.


Lead driver goes off the road
By slatr on 11/11/2009 5:18:56 PM , Rating: 2
So what happens if the lead driver goes off the road? Do the lemmings follow?




By Bruneauinfo on 11/11/2009 5:57:19 PM , Rating: 2
depends on why. But I would hope to avoid an on comming car the lemmings would follow.


By InternetGeek on 11/11/2009 10:04:24 PM , Rating: 2
Easily solved. Have drivers keep their hands on the wheel and mind the wa... oh wait


So how do you get on the motorway then ?
By Landiepete on 11/12/2009 3:47:06 AM , Rating: 2
The motorway I use every morning is very busy, as it's a conntecting road for incoming traffic from the Netherlands.
It is daily fare that you practically can't join in the morning, because trucks are head to tail for miles.
It requires a combination of situation awareness, throttle control, courage and blind luck to get on it it, and thread yourself through the trucks to the left lane, where trucks are not allowed to drive.

So how does this system work then ? If there's too much space between the trucks, the benefit is lost. If the distance is reduced, and there's a 5 mile road train thundering down the road, you have no chance in hell of being able to join the train.

So how do they solve this ?




By heulenwolf on 11/17/2009 10:41:46 AM , Rating: 2
Agreed, merging onto busy roads is challenging enough for drivers. If you have a short merge area and a long train of vehicles is currently taking up the lane you would merge into, you have to stop. Stopping is about the worst thing you can do for traffic flow in busy areas. Perhaps the trains would only operate on roads with enough lanes that they would never be in a merge lane or a passing lane (3 lanes and up). In the US, using the HOV lanes present around many big cities may be a good fit.

Unfortunately, the potential for disastrous pile-ups seems high. Not all vehicles in a train stop at the same rate. If someone jumped in front of the train or, in the HOV example, tried to merge in between two to join one of them you could have a multi-car pile-up where you otherwise may have only had two cars collide. I would guess that part of the "automatic steering system" would also need to be an automatic throttle/break system as well as collision detection and avoidance systems. If the vehicles are close enough to gain a fuel efficiency boost, they're too close for drivers to see what's coming or to react.


By frobizzle on 11/12/2009 8:17:38 AM , Rating: 4
quote:
drivers in the road train would be able to take their hands of the wheel and sleep or eat while they are driven to their destination in their own vehicle.

And this is different from the way most people drive today how??




It might work!
By MrWho on 11/13/2009 2:59:32 AM , Rating: 2
If Volvo is onto it, it might actually work!




"This is from the DailyTech.com. It's a science website." -- Rush Limbaugh









botimage
Copyright 2009 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki