A Lower Cost Grade Crossing Protection System

[BaltACD]

charlie hebdo

You are right. An electronic signal from PTC loco to appropriate vehicle systems would be more effective than blowing its horn, which is an archaic, 19th c. method.  

Safety is more important than "business intelligence" which is often an oxymoron, much the same as PSR.

So - blowing the horn will activate what ever electronic counter measures become STANDARDIZED, just like blowing the horn activates the ditch lights and the bell.

Now how high and wide and what should be the strength of the electronic counter measures 'beam' be?  

[charlie hebdo]

You are right. An electronic signal from PTC loco to appropriate vehicle systems would be more effective than blowing its horn, which is an archaic, 19th c. method.  

Safety is more important than "business intelligence" which is often an oxymoron, much the same as PSR.

[Overmod]

charlie hebdo

As semi- and fully autonomous vehicles become more prevalent, IT and AI folks should be able to make crossings safer through the vehicle.

Although it should be clearly recognized that even a level 4 vehicle without external sensor fusion or inputs might suffer precisely this sort of accident in this way.

Perhaps sensitive microphone arrays would provide earlier directional warning of 'train noise', but no arrangement of cameras would have picked up motion 'through the trees' with sufficient lead time.  Nor could the systems providing vehicle control give a response time faster than the laws of physics permit.

A simple tie-in with PTC information or cheap roadside cameras with better sightlines solves this... when everything is working perfectly.  I would note that the original Carnegie-Mellon approach of tying into GIS to optimize energy consumption and navigation lends itself to relatively secure grade-crossing alerts that won't compromise business intelligence (or necessarily facilitate expedient harvesting) but this goes away if onboard-stored GIS is in use.  I suspect there was limited high-speed wireless data access in that accident location, and while it might be practical to implement some kind of femtocell coverage at these sorts of crossing, I suspect it could not be cheap and might be difficult to make robust or theft/vandalproof.

Of course a properly-designed level-4-autonomous vehicle would take better note of signage and, if properly programmed, recognize the crossing from 'tercom' and use properly conservative action to negotiate it... and realize whether its vehicle will high-center or stall in the effective foul zone.

[charlie hebdo]

As semi- and fully autonomous vehicles become more prevalent, IT and AI folks should be able to make crossings safer through the vehicle.

[tree68]

BaltACD

Have several videos of deer that have been browsing around my driveway and back yard - the back yard is enclosed in your everyday chain link fence.

Story in a book wiritten by a local author:

Two fellows were bragging about their hunting dogs.  One claimed his could follow a year's old track.  Of course, the reply was, "show me."

So the first fellow gives his dog the command to seek out a track.  Soon the dog yelps and sets off on a run across a field.  About half way across the field, the dog suddenly jumps straight up in the air, then resumes tracking.

"See, told you!," says the dog's owner.

"What does that prove?"

"Some years ago there was a six foot page wire fence there..."

A local farm here raises deer - their fences are plenty tall.

[BaltACD]

Erik_Mag

 

SD70Dude

You could fence off the road and build grade separated wildlife crossings.

There's a wildlife overpass crossing US-101 between Augora and Thousand Oaks. I'd also shudder at how much it would cost on a national basis, though I would think the installations would most likely be where "wilderness" areas existed on both sides of a major highway.

Wildlife is everywhere and while you can build structures for them to cross highways and railroads - you have difficulties in 'instructing' them to use those structures; at least without  serious fencing.

Went to the hardware store yesterday afternoon - 'startled' a doe and two fawns less than a mile from my house.

Have several videos of deer that have been browsing around my driveway and back yard - the back yard is enclosed in your everyday chain link fence.

[Erik_Mag]

SD70Dude

You could fence off the road and build grade separated wildlife crossings.

There's a wildlife overpass crossing US-101 between Augora and Thousand Oaks. I'd also shudder at how much it would cost on a national basis, though I would think the installations would most likely be where "wilderness" areas existed on both sides of a major highway.

[Overmod]

Well, there I thought I was defending the point he was making, only to find I seem to have misconstrued what he actually thought he was meaning to say at the time.  (/s)

I do think it is likely that many, perhaps most people will assume that a road crossing will have lights and gates on an active line, and tend to 'stop, look, and listen' less at passive crossings because if a train were likely to be there, the government would have mandated the familiar 'active railroad crossing' paraphernalia.

In my opinion a good low-cost system needs to reinforce the necessity to STOP, LOOK, LISTEN at any crossing -- and to note in its signage when it is across a busy or highly active ... or high-speed... line.  (The Milwaukee, along with those famous 'slow to 90mph' signs, had a somewhat terrifying notice about 'high speed trains' on some of the unsignaled crossings...)

Part of the long-term solution is to do as the Canadians are doing with their 'attention alert' tones (and Holley Rudd did successfully with the alternate-flashing crossing signal): develop signage,approaches, etc. that reinforce public understanding that one is vigilant and thoughtful at any specially marked crossing, as a subconscioudly active response.  Terror and mandates are a poor, even though somewhat effective, way to build that kind of instinctive compliance.  Alternatives welcome.

[Euclid]

I said that active crossings make passive crossing more dangerous.  In other words if you made all active crossings into passive crossings, the crossings that were passive all along would suddenly be less dangerous.  But this is just to demonstrate a principle, and not something I suggest doing.  In fact if it were actually done, it would made the passive crossings safer, but make the previously active crossings more dangerous. 

 

I started to explain this in a reply to Charlie, but I think it got too confusing, so I took it down for some remodeling.  It is not an easy thing to explain.  But the basic idea is that if all crossings suddenly were passive, drivers would no longer be lulled into a false sense of security by the automatic protection of active crossings. 

 

But the point was to be directed at the idea of making passive crossings safer by adding active protection to them, and to do this at a lower cost than that of a typical new active crossing installation. This is what Greyhounds was suggesting in the other post.  This is also what the Minnesota project was attempting to do. I will explain this better later. 

[charlie hebdo]

Overmod

charlie hebdo: So logically you favor unprotected crossings and intersections, ban driver assists?

Overmod: That doesn't follow logically from what he way saying at all.

Mind reading Euclid?  Really?

Actually I was being facetious.

I do like your lower-cost crossing idea and large fines.

[tree68]

Overmod

I already see people thinking that crossings without gates are less important than those with half-gates, and those in turn less serious than full four-gate crossings with little arms for the sidewalks and baffles for the bike traffic.

The crossing in question has an interesting inversion - rather than a less-used rail line on a busy highway, which would have lights and gates in most cases, we have a little used road on a busy rail line.  

And most of that limited number of vehicle crossings had probably, for years, be local folks familiar with it.  That concept has been alluded to here.

It will be interesting to see what the final solution to this particular hazard will be.

[Overmod]

charlie hebdo

So logically you favor unprotected crossings and intersections, ban driver assists?

That doesn't follow logically from what he way saying at all.

What he was pointing out is an example of what right-wingers might term 'nanny-state enablement': the idea that any real railroad crossing would have 'government' signs, lights, and gates and that if it were important there would surely be lights. 

And quite probably the idea that if a train actually hits them at an 'unprotected' crossing they can sue for large money damages because someone should have given them warning.  I already see people thinking that crossings without gates are less important than those with half-gates, and those in turn less serious than full four-gate crossings with little arms for the sidewalks and baffles for the bike traffic.

I think Jeff noted another problem: People are paranoid about police enforcement of things like 'California roll' stops, but have the perception that they wouldn't be ticketed for running a crossing because that's a private company's affair.  I confess that I accept that such mentality is likely to exist, and that it is somewhat behind my preferred alternative of sensor-fused cameras, multiple-hundred-dollar fines and suspensions where justified for repeat offenders -- in fact, all the expedient legal enforcement machinery built up for "drunk drivers" applied to a similarly dangerous social concern.

[tree68]

rdamon

They just need to eliminate the forest/road grade crossings.  

Would that be like the person who called up the highway department and asked them to take down the "deer crossing" signs, as the deer crossing there were a problem...

https://fox4kc.com/news/offbeat/woman-misunderstands-deer-crossing-signs-calls-radio-sation-wants-them-moved/

 

[SD70Dude]

You could fence off the road and build grade separated wildlife crossings.  Most of the Trans-Canada Highway through Banff National Park is like this, and it seems to work.

I shudder to think of how much it would cost on a national level.

[rdamon]

They just need to eliminate the forest/road grade crossings.  

[tree68]

alphas

In a lot of PA, the 2nd most cause of car accidents is deer either running into cars or being hit by them.

I don't know the numbers for NY, but I'm sure they are similar, espcially in northern parts of the state.  I almost got one a couple of nights ago myself.

I was out with my fire department the other morning for - (wait for it) - a car/deer collision.  Folded the hood up pretty good.

The sheriff's deputy who responded said it was the fourth or fifth they'd been to during their midnight shift.

[alphas]

 

"When I had my Dodge Daytona I got hit by a deer that couldn't stop on wet pavement.  It was the 3rd of three that ran across my route on a 40 MPH roadway - the first two made it, the 3rd one applied his brakes in full service but his hooves didn't have any grip on the wet pavement and it took out my passenger side mirror.  Looking at my rear view mirror I saw it shake its head and stumble along in the direction of its friends."

 

In a lot of PA, the 2nd most cause of car accidents is deer either running into cars or being hit by them.   In some of the most rural counties it can be the #1 cause.   Around where I live, deer help keep the auto repair shops pretty busy.  I had one run into the side of my car, cause $3,500 of damage, and ran off without showing any sign of injury.    While my car was being fixed, I was riding in my girl friend's car in a built-up area only to have a massive buck jump a guard rail and come down on her hood.  He immediately took off, again jumped the guard rail without any problem, ran away between the buildings, and caused $3,500 of damage.   Fortunately, deer vs vehicle accidents in PA fall under the Comprenhensive Auto Insurance coverage so they don't have a claims impact on your premium.      

[Euclid]

.

[jeffhergert]

One early morning some years back, we were approaching a railroad crossing on a double tracked main line.  We were going slow because of a temporary 25mph slow order.  About 150 feet south of the crossing was a 4 way stop.  The north/south street was the busier route, the east/west was a stretch of the old Lincoln Highway, still paved with bricks.  (Since this happened, it's now a 2 way stop, stop required for people on the old highway, when it's not closed to traffic.) 

I noticed a northbound car.  It stopped at the 4 way intersection, with nothing in sight.  The crossing has gates and lights and they were working, the gate fully lowered.  After stopping, the car drove up to the crossing and proceeded around the lowered gates.  

I have a question for this driver.  Not why he/she risked his/her (probably his) life by going around the gates, he/she saw us and assumed the gates were down because we were slowly approaching.  (It would've been a surprise if the gates were down because a faster train on the adjacent track had been coming.)  I want to know why this driver stopped at the 4 way stop with nothing in sight.  

I'm not against having active lights and gates.  But I cringe everytime a vehicle is hit at a crossing equipped only with the crossbucks and the media calls it "unprotected."  The crossing was protected, the crossbucks mean something.  A crossing like that is more protected than many rural county road intersections, some of them quite busy.

Now for something completely different.  We have deer whistles on our vehicles.  The instructions say they work over 30 mph.  I've heard they work and don't work, and have had some experiences both ways.  I think if the deer is just milling about next to the road, eating in the ditch, thinking about crossing, etc., that the whistles get their attention and alert them to the vehicle.

If the deer is on the run, spooked by something, running with the herd, etc., the whistles aren't going to have any effect.

I had a young adult (teenage equivalent?) deer once come out on the tracks.  I blew the horn a few times and the deer charged at us.  It lost.

Jeff   

[charlie hebdo]

So logically you favor unprotected crossings and intersections, ban driver assists?

[Euclid]

The least possible protected grade crossing has only the protection of being identifiable just by the appearance of a crossing deck built across the roadway.  With such a crossing, the only protection is the driver’s awareness of the crossing and his/her wariness toward it in watching out for approaching trains.

 

From that point, every bit of added protection reduces driver wariness.  In other words, drivers will naturally depend on the protection and lower their own perception of danger accordingly.  For instance, they will pass through an inactivated active crossing without looking both ways.

 

So, it follows that every active grade crossing today makes every passive grade crossing more dangerous than it would be if all crossings were passive.  This is because drivers have become habituated to relying on the maximum protection of the flashing lights, bells, and gates of active crossings, which lowers their wariness accordingly.  Then with that habitual state of lowered wariness, they apply that to passive crossings.  Primarily they forget the need to look for trains.

 

[BaltACD]

tree68

 

BaltACD

Have they tried Deer Whistles on locmotives? 

I've hit a couple of deer with locomotives...  And a few other animals as well. 

They get spooked and don't know which way to run (just like on the highway).  Sometimes they go the correct way, sometimes they get hit.

A contractor was driving our pumper for some maintenance.  On his way, a deer crossed in front of him.  It was airborne, taking out a windshield wiper, among other things.  On a custom pumper...  The deer was every bit of four feet above the road surface.

Now back to the regularly scheduled program.

When I had my Dodge Daytona I got hit by a deer that couldn't stop on wet pavement.  It was the 3rd of three that ran across my route on a 40 MPH roadway - the first two made it, the 3rd one applied his brakes in full service but his hooves didn't have any grip on the wet pavement and it took out my passenger side mirror.  Looking at my rear view mirror I saw it shake its head and stumble along in the direction of its friends.

[tree68]

BaltACD

Have they tried Deer Whistles on locmotives?

I've hit a couple of deer with locomotives...  And a few other animals as well. 

They get spooked and don't know which way to run (just like on the highway).  Sometimes they go the correct way, sometimes they get hit.

A contractor was driving our pumper for some maintenance.  On his way, a deer crossed in front of him.  It was airborne, taking out a windshield wiper, among other things.  On a custom pumper...  The deer was every bit of four feet above the road surface.

Now back to the regularly scheduled program.

[BaltACD]

rdamon

Maybe we could leverage something like this.

https://www.cbc.ca/news/canada/edmonton/grizzly-bear-train-banff-jasper-u-of-a-research-1.5358433

Have they tried Deer Whistles on locmotives?

I have them on my truck - don't really know if they work or not - have passed many deer on the side of the road when I do my nighttime traveling.  Haven't seen any react one way or another.

[rdamon]

Maybe we could leverage something like this.

https://www.cbc.ca/news/canada/edmonton/grizzly-bear-train-banff-jasper-u-of-a-research-1.5358433

 

[charlie hebdo]

dpeltier

On the other hand, "autonomous" vehicles that react to visual input from sensors have made enormous progress. Instead of trying to build a low-cost active warning system for railroad crossings, we might get more benefit from developing on-vehicle systems that can see a crossbuck, slow down or even stop if approach sight distance is inadequate, scan for incoming trains, and not go across the tracks until it's safe to do so. If you get make that work, then once it is widely deployed (20-30 years?)

Great idea and it might be a lot sooner than that.

[dpeltier]

tree68

 

 

Euclid

Okay, I do recall that $4,000 target.  

 

 

Couldn't find $4,000.

Did find $250,000, and the estimated $400,000, as well as a documented  $768,000 for a full four quandrant set-up.

The Minnesota set-up was $40,000, with inflation taking that to some $60,000.

One could probably go through the shopping list for Rochelle and figure out a rough cost for two sets of lights and gates, along with the required sensing equipment.

 

As of a couple years ago, a baseline cost for a newly-installed lights-and-gates system on a 24' rural road and a single-track railroad, with no nearby intersections, major hills or curves (in the road), switches, signals, detectors, etc. was around $200 - $250k installed. Projecting backwards, in 2005 that cost would probably have been around $150k. The demonstration project set out to reduce the cost to $15k (around 90% reduction), and wound up costing about $40k. If you read the whole report, you will see that the 3x cost overrun mostly came about because the people writing the spec for the project were, like one or more people on this board, willfully ignorant about the actual requirements of a railroad signaling system.

A few comments or highlights to mention from the report:

1.) The equipment installed included flashing lights, but no gates. In 2022, I don't believe any state is willing to do a new active crossing without gates in most cases. I'm not sure how much cost gates would add - there's the motors themselves, extra batteries, extra fail-safe modes to detect and respond to, etc. Plus extra power requirements (see below).

2.) The equipment in the study also included active warning (flashing amber lights) at the advance warning signs (the yellow, circular RXR signs out in front of the crossing). MnDOT uses something similar on rural state highways in advance of a traffic signal ("Prepare to stop when flashing"), but it is not common for railroad crossings. This added some amount to their "per crossing" cost.

3.) It's not clear to me whether the "$40k per crossing" includes the equipment in the locomotives or not.

4.) The report notes that there's a big hurdle to get over for using a system like this anywhere but on a shortline or regional railroad. With the setup they had on the TC&W, they had to install the equipment on a small number of locomotives. If you did this on a Class 1, you would have to install the equipment on basically every  Class 1 locomotive in active service. Doing your FIRST crossing would have a ridiculously low cost/benefit ratio. So you'd have to have something that was had been studied, tested, and fully 100% developed in a number of test environments before you could deploy it generally. The demonstration project didn't do that and made no claim to have done so. It looks like it wasn't a big enough success for anyone to be willing to fund the next step - like a drug that shows some limited benefit in animal studies, but doesn't justify human trials.

5.) The equipment used a 220 MHz band allocated for Intelligent Traffic Systems (ITS) equipment. I don't know if that still exists, whether it's being used for other ITS things, etc. Ideally the same radio bandwidth and radio equipment that is used for PTC could be used, but that creates several more complications (such as insuring that these multiple safety-critical functions don't start interfering with each other). I'm not much of a radio guy, don't know how big an issue that is.

6.) One of the features of the demonstration system was that it didn't require commercial power at each crossing. I think that railroads will take a lot of convincing before relying on off-grid solar panels for safety-critical stuff like this. Power needs would increase significantly if gates were to be added. Fortunately, if you're looking to cut costs by ~50% instead of ~90%, getting power service is not going to be a deal breaker at most crossings.

7.) Lifecycle cost is a big consideration as well, and was not addressed in the report. Most likely that is because MnDOT was the driving force behind the demonstration projecg, and MnDOT (like many states) usually requires the railroad to pay all costs for maintaining the crossing after installation is complete. But if you come up with something that's cheap to install but no cheaper to maintain, and has a shorter lifespan, that could cause problems.

People are definitely working on various kinds of unconventional ("wireless") crossing systems, despite the challenges, because there are some potential benefits. I believe there are three places where PTC-integrated crossings are currently in service: Amtrak in Michigan, Illinois Higher-Speed Rail, and Denver RTD. All three are overlays that extend the crossing approaches beyond what can be achieved with conventional AF circuits in order to allow higher train speeds. None of them has a very good reputation. Other "wireless" approaches, not necessarily reliant on train->crossing comms, are also being explored.

But here's my take: the vast majority of traffic accidents aren't the result of a driver not having access to information, but because the driver is paying attention to something else or makes a bad decision. The whole ITS / CAV business from the 1990's, with "intelligent" cars talking wirelessly to each other and the infrastructure, really hasn't panned out at all. On the other hand, "autonomous" vehicles that react to visual input from sensors have made enormous progress. Instead of trying to build a low-cost active warning system for railroad crossings, we might get more benefit from developing on-vehicle systems that can see a crossbuck, slow down or even stop if approach sight distance is inadequate, scan for incoming trains, and not go across the tracks until it's safe to do so. If you get make that work, then once it is widely deployed (20-30 years?) we could probably even get rid of the signals at quite a few active crossings with no loss of safety. Active warning devices would only be required where sight lines are very, very bad, or on high-volume roads where slowing down as you approach the crossing poses too many other problems.

Dan

[Euclid]

The use of PTC has been mentioned as a means of controlling grade crossing protection systems.  The theory is that since PTC knows where trains are located, it will know when a train is approaching a grade crossing. 

 

As PTC was rolled out, there were proposals to include grade crossing protection activation in the PTC function.  However, as I understand, the primary purpose was not to use PTC as an alternate means of controlling grade crossing protection. 

 

The main purpose was to notify the crew of an approaching train if the crossing was fouled by a stalled vehicle before the vehicle came into sight of the engineer.  This would be an entirely new feature, never before offered.  It would amount to PTC being used to detect track obstructions far in advance of an approaching train such as is done by slide protection and broken rail detections. 

 

So this would take advantage of the ability to warn the engineer of a fouled crossing while there is time to stop short of a collision.  Otherwise, that advantage cannot be taken if the engineer has to wait until the fouling is within eyesight, and thus not allowing enough time/distance to stop the train. 

[tree68]

BaltACD

Can't we just find a maimed former operating department employee and construct a shack on the premises for them to live in and have them flag the crossing for food & WiFi until they die?

Old (really old) school...

[chutton01]

From what I gathered from that report it seems that in effect the active crossing signal would be controlled by signals transmitted from the train itself, as opposed to using line-side detecting gear. I had thought that systems like that had been created by the 1920s (like many other technologies), but whatever.
I will say the art director for that report blew it on Figure 1-1, where the "Passive Control" image on the left has 2 "Active Control" crossing lights standing out along the main-line crossing in the image, while the actual example of a "Passive Control" cross-buck (along I guess a siding in front of the main route) is almost lost in the visual clutter behind the round R/R crossing sign - rather poor composition.
As for figure 4-1. which I guess was created by the director's pre-school children, the clip-out steam loco weathervane coupled with the pinkish London Double-Decker bus kinds of drops the ball on presenting a professional look...

BaltACD

Can't we just find a maimed former operating department employee and construct a shack on the premises for them to live in and have them flag the crossing

Hmm, a number of times so far this year I've had good use for that old 1950s Plymouth ad slogan (trimmed a bit):
"Suddenly, it's 1960..."