I'd guess that when Germany was unified, they would've ended up with (at least) two seperate systems, one from what was East Germany and one from what was West Germany. They might be on one system now, but that wouldn't have happened overnight.
Amtrak says.... About accident of 371-30
- Thread starter Navy 118
- Start date
Help Support Amtrak Unlimited Discussion Forum:
- Status
wayman
Engineer
I'd also guess--and I mean guess, I've done no research here--that there are stronger similarities between the SNCF, SNCB, and NS than there would have been without WWII, given that substantial railroad infrastructure was destroyed in the war and rebuilt with American assistance. (In the Netherlands, the entire country was electrified by the late '30s, but due to war damage, much of the country ran on (USRA, largely) steam again from 1945 into the mid-'50s. I imagine infrastructure damage to their signal systems was up there with their electrification, but again, this is all pre-coffee thinking.)
Still, I'm sure there are substantial differences in spite of any similarities that may have been introduced, due to (a) attempts to replicate to some extent pre-war signaling systems, and (b ) fifty years after the war for the countries to develop separately.
Last edited by a moderator:
R
rdj
Guest
Re accuracy of GPS, when I use it in my airplane, I can see which taxiway I am on at a congested, large airport. So you would think it could be modified for rail service, at least as a safety backup.
Joel N. Weber II
Engineer
I think the typical wingspan on a smaller airplane is close to 50 feet. So it wouldn't surprise me if 50 foot GPS accuracy does indeed easily tell you which taxiway you're on, whereas you can easily find multiple tracks in the space of 50 feet.
But I'm also not sure that knowing which track you're on would be key to avoiding this accident; the other key factor is knowing what else is on the track ahead.
The GPS on your airplane does not by itself tell you where all the other airplanes are, and I'm not aware of any readily available technology that will show you where all the airplanes on the ground at an airport are on a display in your airplane.
And there's still plenty of airspace in the US where, no matter how fancy the equipment you have operating on the transponder frequency, you still need to look out the window to avoid VFR flights, even when you're flying IFR, unless you're in IMC. I'm not sure how a railroad signal that says to go 15 mph, or slower if you need to go slower to have adequate visibility, is much different.
R
rdj
Guest
It seems the Alaska RR is already using GPS for colision avoidance. Here's from a recent article:
To help monitor and manage its trains in this dynamic environment, the company recently implemented the first phases of a Collision Avoidance System (CAS), based primarily on real-time GPS positioning and VHF data radio communications. The system is designed to prevent train-to-train collisions, protect track maintenance forces, and enforce speed limits.
To help monitor and manage its trains in this dynamic environment, the company recently implemented the first phases of a Collision Avoidance System (CAS), based primarily on real-time GPS positioning and VHF data radio communications. The system is designed to prevent train-to-train collisions, protect track maintenance forces, and enforce speed limits.
Last edited:
R
rdj
Guest
And the use of GPS gets even more interesting. UP is using it to ID BAD RR TIES. If they can do that, they sure as heck can ID the location of a freight train. Here's from another article:
The Union Pacific Engineering Systems Group deployed a system for railroad tie
inspection to be used by the "System Tie Spotters." The solution included Symbol 8100
handheld computers with snap-on modems and LinksPoint Bluetooth GPS. With the
solution, the spotter walks the tracks and uses a specially-designed “clicker” to count
bad ties. Each “click” generates a record on the handheld computer that is associated
with a precise location captured from the LinksPoint Bluetooth GPS receiver and
transmitted to the handheld computer using Bluetooth short-range wireless technology.
The tie inspection software used on the handheld was designed in-house with
the help of LinksPoint’s GPS application programming interface (API) and engineering
support to access GPS position data. Specific components for synchronizing data with
Union Pacific’s ERP system were also developed by the Engineering System Group.
The project was initially deployed as a pilot and based on its success was rolled out to
the full team of Spotters.
When deployed in the field with the team of 20 Spotters, some results were
immediately apparent. Spotters could now walk the tracks and use the clicker to record
GPS location data for each bad tie. At the end of each assignment, the Spotter uses
the handheld’s modem to dial into the railroad’s planning system to automatically
upload bad tie location data collected in the field. The planning system uses this in the maintenance cycle to plot specific
locations for the unloading of replacement tie bundles. GPS on the replacement tie train used the plotted locations to
automatically drop the bundles for use by the repair crews.
Benefits
The Union Pacific mobile solution is very successful in that it provides benefits far beyond simply
automating a process. It eliminates the use of paper recordkeeping and additional data entry. It transforms
The Union Pacific Engineering Systems Group deployed a system for railroad tie
inspection to be used by the "System Tie Spotters." The solution included Symbol 8100
handheld computers with snap-on modems and LinksPoint Bluetooth GPS. With the
solution, the spotter walks the tracks and uses a specially-designed “clicker” to count
bad ties. Each “click” generates a record on the handheld computer that is associated
with a precise location captured from the LinksPoint Bluetooth GPS receiver and
transmitted to the handheld computer using Bluetooth short-range wireless technology.
The tie inspection software used on the handheld was designed in-house with
the help of LinksPoint’s GPS application programming interface (API) and engineering
support to access GPS position data. Specific components for synchronizing data with
Union Pacific’s ERP system were also developed by the Engineering System Group.
The project was initially deployed as a pilot and based on its success was rolled out to
the full team of Spotters.
When deployed in the field with the team of 20 Spotters, some results were
immediately apparent. Spotters could now walk the tracks and use the clicker to record
GPS location data for each bad tie. At the end of each assignment, the Spotter uses
the handheld’s modem to dial into the railroad’s planning system to automatically
upload bad tie location data collected in the field. The planning system uses this in the maintenance cycle to plot specific
locations for the unloading of replacement tie bundles. GPS on the replacement tie train used the plotted locations to
automatically drop the bundles for use by the repair crews.
Benefits
The Union Pacific mobile solution is very successful in that it provides benefits far beyond simply
automating a process. It eliminates the use of paper recordkeeping and additional data entry. It transforms
George Harris
Engineer
"Apples and oranges" comparison. The Alaska Railroad is a single track railroad operating at relatively low speeds in a very remote territory for the most part. One of the key words here" Single track" 50 feet or even 200 feet would be good enough. Not to mention the following sentence not quoted sounds like it was taken from promotional literature by the manufacturer or supplier. The reality may be somewhat different. Nothing says this is a stand alone system. There probably still has to be train to train or train to dispatcher communication to ensure that at meets the train that is supposed to be in the siding really is in the siding.
More sales spiel. Again, locating ties within 50 feet, considering that the ties are also marked, is an entirely different animal from locating a train within 6 feet. (Track centers can be as low as 13 feet, so if the accuracy of your system can not place you to within less than half that distance there is too much uncertainty left.)
Sorry, bu now I have heard of too many wonderful solutions developed by someone that is in reality clueless concerning the nature of the problem they think they have solved.
George
Last edited by a moderator:
transit54
Conductor
My understanding of GPS technology is that it can be accurate within 3 meters. Straight GPS is 50 feet or so as far as accuracy, but provided that you have a lock on a WAAS satellite, you can be accurate within 3 meters (WAAS basically corrects the GPS signals for a greater degree of precision). However, there's a few factors that are not being considered as far as GPS and railways.
For one, GPS needs line of sight to the sky get a decent, accurate signal. This works great for planes. However, consider a lengthy railway tunnel. Suddenly, your train, previously accurate to three meters, is nowhere to be found! Imagine the safety implications. Also, consider lines with substantial foliage overlaying the tracks. This will cut down on the accuracy of the GPS signal, now bringing you back down to your ~50 ft or so accuracy, which has been noted as unacceptable.
That being said, my understanding is that GE (and possibly EMD) now builds an optional package into a lot of their locos that monitors diagnostics and reports such information back to Erie, as well as including a GPS receiver so the location of all the monitored locomotives can be tracked for service reasons. I believe Trains had an article on it in the last year or two - I'd try and dig it out but I'm away from home for about a week.
GPS definitely has a role in railroading, but I wouldn't equate its role in aviation to its potential when applied to trains.
For one, GPS needs line of sight to the sky get a decent, accurate signal. This works great for planes. However, consider a lengthy railway tunnel. Suddenly, your train, previously accurate to three meters, is nowhere to be found! Imagine the safety implications. Also, consider lines with substantial foliage overlaying the tracks. This will cut down on the accuracy of the GPS signal, now bringing you back down to your ~50 ft or so accuracy, which has been noted as unacceptable.
That being said, my understanding is that GE (and possibly EMD) now builds an optional package into a lot of their locos that monitors diagnostics and reports such information back to Erie, as well as including a GPS receiver so the location of all the monitored locomotives can be tracked for service reasons. I believe Trains had an article on it in the last year or two - I'd try and dig it out but I'm away from home for about a week.
GPS definitely has a role in railroading, but I wouldn't equate its role in aviation to its potential when applied to trains.
battalion51
Engineer
Most engines these days already have GPS systems on them for tracking purposes. The beauty of the railway is that you can know the spots of every train with the electrical currents that run through the rails. With cab signaling accidents like the Chicago incident could have been prevented with a system like ATS or ACSES.
Dutchrailnut
Conductor
ATC not S or ACSES, is fine but the drawback is all locomotives leading in such a territory need to be equipped. in Area's like Chicago which is hub of American railroading it would require about 16 000 locomotives plus the tracks around chicago to be equipped.
with todays run trough power, fleets are no longer captive.
and with ATC/ACES you can't just equip passenger trains and let freight trains run unprotected, so its all or nothing, so who is gone pay for the $50 000 instal per locomotive of euipment plus labor and who would pay for the wayside equipment.??
Joel N. Weber II
Engineer
They'd probably end up being LCD, not LED.
And http://www.avweb.com/news/sayagain/191072-1.html seems to demonstrate that even when you give one person a display showing all the vehicles in an area, fatal collisions can happen.
I'm also not sure how you'd design this GPS sytem's screen to be easier to interpret than standardized wayside signals would be, if you could get all the railroads to standardize on one set of signals. You probably don't want it displaying all of the parallel tracks, because then if the engineer looks at the wrong track, he'll/she'll take the wrong action. (Unless you can find a good way to highlight the track that the train is heading down. Color might work if colorblindness (or the particular form of colorblindness that would matter) is considered disqualifying for a job as an engineer.) So you end up wanting something that primarily shows how far away the next train on the current track is, which oddly enough the current signaling system already gives you a rough approximation of.
Another issue to keep in mind is that if you only put this system in the locomotive at the front of the train, you need to somehow have the system have some idea about where the back of the train is. Since the whole point of building this GPS system is that someone assumes engineers can't be trusted to read signals, it might be safest to assume that they can't be trusted to enter the length of their freight train correctly, either. For that matter, maybe they can't even correctly enter whether the locomotive is on the east or the west end of the train.
So if you assume that trains take up a couple miles centered on the locomotive, if you can figure out which of those tracks on that 50 foot wide right of way the train is on somehow (maybe via a more sophisticated track circuit that goes beyond the current cab signaling that tells you which track number you're on as well as what the signal indication for your block is?) you might be right back to having an indication that says you need to be able to stop within 1/2 your field of vision, not to exceed 15mph, whenever you're near another train. (But if you more or less need to put in all the track wiring for cab signaling anyway, why not just rely on traditional cab signaling to do automatic stop? I guess the 15mph enforcement isn't going to enforce the 1/2 field of vision thing, but if you're building a fancier track circuit than we have now anyway, why not provide additional indications for 5mph and 10mph or whatever, and enforce a speed that is appropriate for the lowest visibility in the block with worst case brakes?)
Another thing to keep in mind is that railroads have traditionally been distrusting of transistors when they build their signaling systems, because transistors can fail in either direction. Apparently mechanical relays that rely on gravity tend to be highly unlikely to fail in one direction, and the system can be designed so that if a relay fails in the direction favored by gravity, the trains will stop instead of colliding. And I'm not aware of any GPS reciever that doesn't rely on transistors.
wayman
Engineer
Can't FREDs (or whatever they're calling them these days) help with the end-of-train location? Between those and the latest hotbox detection the train has passed, you can get a more-or-less accurate count of axles and a reasonable estimate +/- a few hundred feet of length.... Of course, that's the difference between a collision and a near-collision in some cases!
I'm certainly not advocating GPS or the like--I agree, it just wouldn't work for trains. Just thought this might be an answer to the train-length question.
Joel N. Weber II
Engineer
Given that this Amtrak accident that we're thinking about happened near a yard, if I remember correctly, you'd need to make sure that trains pass a hotbox detector as soon as they got out of a yard, but installing hotbox detectors next to the entrances to yards may be possible. Then again, if they need significant space next to tracks, there may not be that space without investing a lot of money in reconfiguring the yard to make that space available.
If the FRED also had a GPS and radio, and you could reliably know which FRED is ultimately coupled to which locomotive, you could get train length data there (though it's still not a trivial problem if the track has curves, for which your GPS system needs a detailed track database) and I'm not sure if the total number of feet of slack in the couplers you get in a 100 car train is significant. But I'm not sure how you make sure that the FRED and the locomotive get matching ID numbers or whatever it is that they need more reliably than you get people figuring out that red over yellow means 15mph or 1/2 field of vision.
Also, isn't it the case that frieght trains sometimes operate with locomotives in the middle and/or back of the train? How do you make sure that the GPS based system never gets dangerously misconfigured when that happens?
If the FRED also had a GPS and radio, and you could reliably know which FRED is ultimately coupled to which locomotive, you could get train length data there (though it's still not a trivial problem if the track has curves, for which your GPS system needs a detailed track database) and I'm not sure if the total number of feet of slack in the couplers you get in a 100 car train is significant. But I'm not sure how you make sure that the FRED and the locomotive get matching ID numbers or whatever it is that they need more reliably than you get people figuring out that red over yellow means 15mph or 1/2 field of vision.
Also, isn't it the case that frieght trains sometimes operate with locomotives in the middle and/or back of the train? How do you make sure that the GPS based system never gets dangerously misconfigured when that happens?
battalion51
Engineer
I was suggesting to equip everything. While expensive the safety benefits would be huge. And I did mean ATS for Automatic Train Stop, I believe that's an old Santa Fe/BNSF system...
M
meatpuff
Guest
Look, there are plenty of ways to do this, and there are more lessons from the aviation world. In aviation, you have the transponder, which most every aircraft carries, that when interrogated by radio on a frequency known to both parties sends back a radio signal; the content of the signal tells the "name" of the aircraft, and the time delay is used to calculate the distance from the interrogator. Usually airplanes carry only the transponder, not the interrogating equipment. The transponder is trivially cheap, and is installed on any forty year-old aircraft that you can buy used for $40,000. It would be simple to put interrogating equipment and a transponder on each locomotive, so each loco could know the real-time distance of each train in front and in back of him for tens of miles. No, you couldn't get it off the shelf, but a toned-down prototype system would make a fine senior project for electrical engineering students at a decent engineering school. This is 1960s technology; in volume you could add it on to each $2 million dollar loco for probably the cost of the engineer's radio.
There are plenty of issues like tunnels, having a system to tell between adjacent tracks (you could have the transponder send this info along with the "name"), not to mention actual efficacy in preventing accidents, etc. It would be simple enough to add a computer, nay, microcontroller, to take the loco's current location and speed together with those of others' nearby and have a voice in the cab shout warnings when a collision is imminent. Similar (but way simpler) to the TCAS systems that come on most any small jet costing a few million dollars.
The point here is just that a collision avoidance system doesn't depend on "modern technology" such as GPS at all. (In fact, note that no air traffic control system in the USA uses GPS technology at all in tracking aircraft and avoiding collisions.) For whatever reason, this application never caught on with railroads. A few observations:
1) The old saying goes that all FAA safety rules are written in blood, i.e. each new rule was made after someone gets themselves or a planeload of passengers killed. The pain point is much higher for aircraft, with lots of very public and embarassing deaths when an airliner goes down. Passengers are such a small part of what's hauled on the rails, and the media doesn't care in the usual case where you tip over some hoppers full of coal when trains collide/derail/etc. Add this to the reality that the danger of collision is remote compared to hitting a truck at a crossing, and you see why freight railroads aren't interested.
2) OBS is sure correct about error in human actions/systems. The guy on 370/371 was "told" (based on his training) that he should have been going 15 MPH instead of 40 MPH. A computer voice intoning "E-brake...e-brake...e-brake..." is maybe more user-friendly, but still no good if the human doesn't follow through!
3) I read Alan describe one time on here how if the Amtrak engineer doesn't blow the horn or manipulate the controls at all for two or three minutes, the train will alert him, and then stop the train dead on the tracks if the engineer still doesn't respond. I was actually surprised Amtrak locomotives had this amount of automation. So I suspect Amtrak and commuter railroads might actually be interested in such a technology; but it does less good if you can't "see" the majority users of the rails: freight trains.
-meatpuff
Last edited:
The aircraft transponder system operation would be difficult to implement for rail and wouldn't provide the same degree of protection anyway; It works so well for aircraft because, since the aircraft are at altitude, a fixed radar on the ground has line-of-sight communication with up to a hundred or so aircraft at a time, for ten to a hundred miles in all directions, depending only on the altitude of the aircraft. A very rough range estimate (in miles) is the square root of 2 times the square root of the altitude in feet, assuming level ground and no obstructions between the aircraft and the radar site. This covers approximate line-of-sight for curvature of the Earth, but disregards path loss and inverse-square law for energy radiation. A train, on the other hand, runs at ground level. That means communications towers all over the place along the route of the train, with at least two transponder interrogators at each one, operating at all times, and then a data link from each of those transponders to the dispatcher. You also have the problem that since it is not integrated with a radar, you don't know which side of the transponder interrogator the train is on, unless you put TWO interrogators on each tower, or more than two, if there are diverging track paths and multiple routes, near that tower. Each of those interrogators would have to have a very high-gain unidirectional antenna. The aircraft system, since it operates in conjunction with a radar, links the transponder return with the radar return, so it gives you an azimuth, an altitude (with a mode-C transponder) as well as a distance, fixing the aircraft in three dimensions. The train system can't do that without at least two directional interrogators on each tower. Now, yes, there are currently voice radio links in place, but you would still have to make a huge investment in additional equipment to add all those transponder interrogators, the data links, and then the computer and display equipment at the dispatch locations.
Then, once you do all that, what does it get you? Aircraft operate with three kinds of separation, really - horizontal, vertical, and time. And those separations are large - thousands of feet vertically, miles horizontally, minutes in time. Two trains, on two adjacent tracks, running in opposite directions, will pass within six feet of each other. Do that with airplanes and several thousand people running the system will all have a coronary. In the accident we are currently talking about, your transponder system would have told everybody that four trains had just collided. Both adjacent tracks to the Amtrak train had moving trains on them - the one to the left was moving in the same direction, the one to the right was moving slowly in the opposite direction. The one on the same track, ahead, was stopped. The aircraft transponder system would have been spitting out collision warnings for ALL of those trains, for probably twenty minutes, at least, and three out of the four would have been erroneous. You can't just take a system designed for another transportation mode and apply it where it wasn't intended, and expect it to be useful.
Then, once you do all that, what does it get you? Aircraft operate with three kinds of separation, really - horizontal, vertical, and time. And those separations are large - thousands of feet vertically, miles horizontally, minutes in time. Two trains, on two adjacent tracks, running in opposite directions, will pass within six feet of each other. Do that with airplanes and several thousand people running the system will all have a coronary. In the accident we are currently talking about, your transponder system would have told everybody that four trains had just collided. Both adjacent tracks to the Amtrak train had moving trains on them - the one to the left was moving in the same direction, the one to the right was moving slowly in the opposite direction. The one on the same track, ahead, was stopped. The aircraft transponder system would have been spitting out collision warnings for ALL of those trains, for probably twenty minutes, at least, and three out of the four would have been erroneous. You can't just take a system designed for another transportation mode and apply it where it wasn't intended, and expect it to be useful.
VentureForth
Engineer
Very interesting ideas, Meatpuff. GPS can be good to within inches - accurate enough to send a missile through a window. But that's military talk and so enough about that. Civilian GPS use is accurate to within 3 meters - 9 feet. With WAAS and a very high update frequency (much more than the once per second offered by most Garmin handhelds), I think that accuracy can be brought to within 3 feet (one meter). The reliability of that accuracy will always be in question. That's why a handheld can be bought for $100, but an IFR approach rated in dash system costs thousands.
Certainly an in-dash IFR GPS system can be used to line a plane up with the centerline of a runway. Heck, UAV rely on this technology to arrive and land safely. I think that that level of accuracy is available for trains, but again the reliability of that accuracy is very important. In an aircraft, it's not much of an issue to land 13 feet off the centerline of a runway. Not ideal, but you're probably not going to hit anything unless you're in a 747 on a 100' wide runway. That 13 feet in a yard at road speeds just isn't good enough for a railroad.
Block sensors and an Automatic Train Control system is hard-wired and can be extremely reliable. And, it doesn't have a latteral sensitivity.
The transponder issue, also, is a very important consideration. It would be near impossible to have a transponder incorporated into a FRED and be useful in a yard. Too many curves to be useful. Trains on parallel tracks would be setting it off all the time. That's why TCAS wouldn't really work in a yard environment. Out on the open road? Perhaps, but again how can you REALLY tell if that train ahead of you is on the mainline or 14 feet over to the left?
I'm not against technology at all. I think that a combination of technology and of old fashioned procedures can go a long way. In Japan, whenever a train goes from one system to another, usually the controls are taken over by a person trained in that section. Keep the engineers relatively local and on a system they are familiar with. Six different rule books for a 2 hour journey is ridiculous.
Certainly an in-dash IFR GPS system can be used to line a plane up with the centerline of a runway. Heck, UAV rely on this technology to arrive and land safely. I think that that level of accuracy is available for trains, but again the reliability of that accuracy is very important. In an aircraft, it's not much of an issue to land 13 feet off the centerline of a runway. Not ideal, but you're probably not going to hit anything unless you're in a 747 on a 100' wide runway. That 13 feet in a yard at road speeds just isn't good enough for a railroad.
Block sensors and an Automatic Train Control system is hard-wired and can be extremely reliable. And, it doesn't have a latteral sensitivity.
The transponder issue, also, is a very important consideration. It would be near impossible to have a transponder incorporated into a FRED and be useful in a yard. Too many curves to be useful. Trains on parallel tracks would be setting it off all the time. That's why TCAS wouldn't really work in a yard environment. Out on the open road? Perhaps, but again how can you REALLY tell if that train ahead of you is on the mainline or 14 feet over to the left?
I'm not against technology at all. I think that a combination of technology and of old fashioned procedures can go a long way. In Japan, whenever a train goes from one system to another, usually the controls are taken over by a person trained in that section. Keep the engineers relatively local and on a system they are familiar with. Six different rule books for a 2 hour journey is ridiculous.
Last edited by a moderator:
wayman
Engineer
So, so true! I only started to understand this on the day of the Pere Marquette accident, which was coincidentally my first day as a volunteer trainman on the West Chester Railroad (volunteer-operated heritage railroad operating on eight miles of no-longer-used SEPTA track). Since it's still SEPTA-owned track and is connected to the national system, engineers and conductors are on the clock, and everything is run by-the-book under NORAC. Though I'm not responsible for learning NORAC backwards and forwards just to start as a volunteer train crewman, it's obviously a good idea to become familiar with it (and eventually to know it intimately).
If I were a career engineer or conductor or foreman, etc, I could certainly learn this down cold, making a priority of it and knowing my job--and far more importantly, my and others' lives--depended on that. (And many WCRR conductors and engineers are current or former career railroaders; others are extremely dedicated volunteers; everyone maintains certification through the FRA channels.) But seeing NORAC and seeing exactly what can happen when multiple rule books must be known equally thoroughly really gave me pause. If the rule books were completely different across the board, that would be one thing--a lot more to learn, but much easier to keep separate in your head. But they're not completely different--they're (as I gather) obviously different in a few places (like positional versus colored signals), similar in some places, and subtly different in other places, but those subtle differences are absolutely critical. And even the "obvious" differences can have subtleties--maybe in addition to "medium approach" looking obviously different with two completely different signal systems it also means something slightly different. (I don't know, but I assume that sort of thing is true in places.) That's much harder to learn, and much easier for even a seasoned engineer to make a minor mistake with.
It's not like having to know German and French, which is a lot to learn but they're extremely different; it's like having to know German and Dutch. But when you slip and use the German word for something in the Netherlands, people just smile (or if they're old enough to have memories of the German occupation, maybe they glare at you briefly before realizing that's impolite); when you slip and use the wrong railroad rulebook for a single rule, people can die.
It gives me a far, far deeper respect for train crews, especially those who do have to operate under multiple rule books regularly.
But while I agree with the idea that crews should switch off so they focus entirely on a single rule book, how could that work when the ten miles on a route into or out of a major terminal can span multiple books? Given the economics (freight railroads will never agree on a single national rulebook) and practicalities (switching train crews every couple hours is one thing, switching a few times within twenty minutes is another), it seems like Amtrak is really stuck with having to have cream-of-the-crop train crews who know multiple books.
Last edited by a moderator:
M
meatpuff
Guest
Well, in my defense I already spent about half of my original post about this explaining why it isn't any good
So I'm not saying the freight railroads should do something like this at all. But for sport's sake:1) I acknowledge the line-of-sight thing is a great plus for aircraft and makes it work really well over long ranges, and that buying expensive equipment to make this happen is impractical. Fortunately these transponder interrogation and reply signals would be tiny information-wise and require very little BW. So you have two choices: integrate it with voice is the best one. You could just throw a little signal right on the same voice channel outside of audio range and if you design everything right have the repeaters that are in place handle it, and it just works without buying any new equipment. Choice two is to find a secure way to do it on a lower frequency so you will have some sky wave propagation and range is no longer a problem.
2) As I said originally, you need a way for the train to know what track it's on (and preferrably speed and forward/backward, more later). It sends this information in response to interrogation. Same as you need aircraft transponders to reply with altitude information for air traffic control radar to work with the best accuracy. I'm not sure of the best way. It would help if I knew more about train signaling technology.
3) This is only for collision avoidance! I'm not proposing at all to make a whole new top-down train dispatching system based on transponders (surely GPS would be a better way to do this). So I'm not worried about any data links, computer screens, or gathering the information in a central location at all. The only entities communicating in this scheme is one locomotive to all others within range. Furthermore, since it's a 1-D problem, we can get away from having to know bearing and angle of elevation like for airplanes: distance is enough! If my computer continuously knows the distance between Alice's train and Bob's train (and of course what tracks we're on), that's enough to predict a collision. Now, we still need to know if that's distance in front or behind to tell if we're about to run into Bob or if he's overtaking us from behind! If we send the current "speedometer" reading along with info about what track we're on when our transponder is interrogated, we need only absolute distance, and can do away with directional antennas and all the rest of that!
Notice the analogy to TCAS, which mediates a discussion only between two aircraft in order to prevent a collision, and does not communicate to any centralized "dispatch".
One way around the "which track" problem is by having a different "squawk" code, for each track, siding, etc. Now if Alice and Bob are approaching each other on a double-tracked mainline on different tracks, their transponder units ignore each other because they're using different codes. One could envision maps of the route showing where the engineer must change squawk codes at certain mileposts as the train changes tracks, similar to how such a map would show slowzones today. You could find ways to automate this...
So once you put all these refinements in place that I have described above, I think you have a collision avoidance system that would work just fine. It's got some more subtle potential problems left that I won't get into. And with relatively old technology and little capital investment beyond an antenna and radio added to each locomotive. Just saying it is feasible.
(Anyone want to start a company with me? HA! :lol: )
Just dont follow your logic, sorry. If you are going to have rulebook changes mid journey, then surely a change at a border is a more reasonable and logical place to have one than some random point between A and B. Your comments about language 'problems' seem a bit confused as well, most Europeans seem to deal with multiple languages a lot better than Americans or us British, taking your argument to its logical ending, then a country that speaks 4 languages should have the worst and most shambolic railway system, but Switzerland seems to disprove your point.
Seems to me American railroads should have it sorted, 1 country 1 language but you ruin it with dozens of rulebooks and multiple changes in how signals should be read in short journey times and all backed up with no form of Automatic train protection. As professional as most railwaymen around the world are, its all going to end in tears. Company dogma and historical precedent dont make for a safe railway.
George Harris
Engineer
I realize that I will not be able to explain it to you to your satisfaction, so I am not really going to try. Suffice to say, yes in a perfect world, we would have one rule book, but we don't. We make the system work, and work quite well, thank you very much, primarily by have competent people doing the job. And do it with a safety record that no one needs to be ashamed about. Our safety record is much better than it looks because all significant incidents generally are quite openly and publically discussed, a situation not prevailing in much of the rest of the world.
Not disputing that, but even competent people make mistakes. Sometimes it all works well despite the system.
- Status
Latest posts
-
-
-
-
-
-
Brightline Trains Florida discussion 2024 H1- Latest: VentureForth
-
-
