GPS Automatic Train Stop

[dharmon]

QUOTE: Originally posted by kenneo

QUOTE: Originally posted by dharmon

I agree. My thoughts are pretty much based on aviation traffic control and military data link expereince not RRs. Now if the RRs need to design a system to locate, designate and target bad guys we're in there..[:)]

[^]Your suggestions have included a redundantcy level that isn't seen in railroading. They won't duplicate all of the CTC that they have in place. What you have mostly been suggesting would work quite well in dark and ABS territories. Stay with us. You have a lot to contribute![:D]

[:)]In my job, redunancy rules[:)] But then again my industry is in the business of efficiently breaking things not fixing them

[jchnhtfd]

Thanks, Eric -- the definitions help a lot, and should help keep us on track, if not coherent.

In my outfit, it was wise never to crticise the cooking -- if you did, you were automatically elected to be the cook until the next dingbat criticised you!

[kenneo]

QUOTE: Originally posted by dharmon

I agree. My thoughts are pretty much based on aviation traffic control and military data link expereince not RRs. Now if the RRs need to design a system to locate, designate and target bad guys we're in there..[:)]

[^]Your suggestions have included a redundantcy level that isn't seen in railroading. They won't duplicate all of the CTC that they have in place. What you have mostly been suggesting would work quite well in dark and ABS territories. Stay with us. You have a lot to contribute![:D]

[kenneo]

We are getting some things crossed up here. First, some definations:

Dark Territory = Routes and other trackage where trains operate without block signal protection and/or authority.

Lit Territory = Routs and other trackage where signals, either automatically or centrally controlled, provide block signal protection and/or authority.

CTC = Generic term for centrally controlled routes and other trackage where block signals provide train movement authority controlled at the central location and displayed in the field by signal indication, where signals provide block signal protection and/or authority, and where some, many or all of the switches are controlled and operated from the central location (power switches), and where all switches have electric switch locks whether "power" switches or "manual" switches.

CTC is the trademark for US&S branded Centralized Traffic Control.

To enter or pass a Control Point, the train must pass at least one Absolute Signal.

ATS = Automatic Train Stop.

ATC = Automatic Train Control

Switch Conditions and States = Three States == Normal (normal route), Reverse (siding or spur), Transit (or In Transit) (when switch points are not lined for either normal or reverse positions and (usually) are in motion between those two states) and two Conditions == Locked and Unlocked.

Meet = where two trains go by each other headlight to headlight, each on its own track.

Pass = when one train overtakes another each on its own track.

Head-on = A meet where both trains are on the same track.

Rear-ender = A pass where both trains are on the same track.

Side colision or Side-swipe = any other crash other than Head-on, Rear-end, or derailment.

Derailment (on the ground) (on the ties) = wreck not caused by a crash.

ABS = Automatic Block Signal System. Used in all lit territories except the Control Point signals (absolute signals or A Blocks). Detected track occupancy controls the signal indication. Of the variations on a theme, we have APB and Automat Interlocking installations as the most common variants.

Of all of the situations that we have been talking about, CTC has all of the items and programming and detectors EXCEPT occupancy of a route segment or track by multiple trains and ATS.

We can make our suggestions as simple or as complicated as we want, but "job one" has been ATS in CTC using GPS and other items such as RFIC. Job two is to do the same in lit territory, but this will be a bit more complicated. Lastly, by far the most complicated, is dark territory.

Part of how to make it work will have been addressed in choosing the systems. It is finished in this step (Step Two).

Then, What are the "show stoppers" in implimenting Steps One and Two. Part of that answer will be answering "Why hasn't it already been done?"

All of the above is a huge load, so by keeping focused on what part we are dealing with at a given point, we can proceed to the others, and possibly reach a reasonable solution.

This mudhops beans are ready. At his place, it is not wise to irritate the cook if you want her to provide yummies for your tummy. More later. (Scary, isn't it!)

[dharmon]

I agree. My thoughts are pretty much based on aviation traffic control and military data link expereince not RRs. Now if the RRs need to design a system to locate, designate and target bad guys we're in there..[:)]

[jchnhtfd]

dharmon -- sounds like we're all on the same track, more or less! Most encouraging (in terms of trains of thought, not trains of say coal!). Occupancy detectors would work instead of the track identifier RFIC system, although it would be necessary to be careful to arrange them so that they weren't fooled by, for instance, an unloaded flat into thinking that that was the end of the train.

I like the idea of a palm pilot type device to throw a dark territory switch and broadcast the resulting status (a switch has three states: normal, unlocked, reverse). Might be an improvement over the old lock system which, as I noted above, can and has been defeated by enterprising youths and sundry other unauthorized folks...

Keep em coming -- but I still want to hear from the dispatchers, engineers, signal maintainers, and 'suits'! This all sounds so promising, yet it hasn't been done -- there must be a reason.

[dharmon]

As far as confict resolution, lets say you have a single track main in dark country. The central control is periodically polling all units (locos) for updates. Eastbound train A's transponder shows that it is traveling at x speed and is 30 miles from a siding. Westbound train B shows that it is 40 miles from the same siding. At some preset distance from each other, for arguements sake the trip distance for single track opposing meets is 60 miles. At sixty miles from each other, if the operator was asleep or not paying attention, the central dispatch/route controller whatever he'd be called gets a conflict alarm on the screen, highlighting the conflicting units. He can direct the subordinate train into the siding and set the switch and signals electronically and in the course of doing so modify Train A's order to take the siding and stop and for Train B to slow. Crews in both cabs could see the order and know what to expect. Train A's crew acknolwdges either electonically or by radio and Train B's crew can see/hear the acknowlegement. If Train A's crew fails to respond, or begins to pass the siding without entering it, central control can issue a stop order to the locomotive (think ONSTAR on steriods). The operator can order Train B's crew to stop before reaching the siding and avoid the collision or direct them to enter it from the other side, whichever is more feasble. If no action is taken on any part to correct the conflict the system could be set to stop both trains at say 10 miles automatically. It might require and awkward reverse move on the main, but that's better than a cornfield meeting. These distances are just made up.

On double track or anyplace where there is GPS positional ambuiguity, an occpancy detector would be required to confirm that Train A was actually on the siding and clear at both ends before the conflict can be cleared from the system. It could be as easy using something similiar to a garage door photo-optic device at either end of the siding to show entry into the siding and clear at the ends once in.

On locals where manual switchs are being thrown, the crew can highlight that swtich on the route map, and show it thrown. The transponder send a signal to central displaying that event on the "big screen". The crew makes the pick up/drop off reenters the main, resets the switch. Then updates the information which is forwarded to central. Until the crew updates the master, central will presume that the switch is still thrown and it would present a route conflict to any train which was going to pass through and was in say 30 miles or so, which would require resolution as mentioned above.

The thing is that this has limitless possibilities. The dude actually throwing the switch, Ed for instance, can do this with a palm pilot sized unit, so the engineer doesn't have to. But it would update all the receivers in the system. Track gang foremen like in Ed's Form B scenario would have a unit to update the track status, so all involved know that he still has the track or given it up.

[jchnhtfd]

I think there is a misunderstanding. The RFID chips proposed here are NOT the ones on the equipment. They are on the track structure, and are used to identify the track. An RFID reader is nowhere near $100K; $1K is more like it, although with hardening to suit it for locomotives it might be more. There would be one on lead locomotives.

The RFID chips proposed here are used to resolve the ambiguity as to which track a train is on. Unfortunately, GPS can't do that reliably. It would be nice if it could, but it simply can't, not for railroad reliability.

I personally don't think that a computer system can or should be used to issue clearances. There are, fundamentally, three situations with which we are dealing: CTC or other central control territory, lit but not CTC territory, and dark territory.

In CTC territory, it is the dispatcher's function to decide who goes where and when, and he or she sets the signals and routes accordingly. The system under discussion here notes whether the signals are being obeyed, and applies appropriate penalty measures if they are not. In lit but not CTC territory, much the same applies, except that the signals and routes are locally controlled. In dark territory, it would be necessary for there to be either a new central conflict resolution capability or distributed (on each train) conflict resolution capability. Keep in mind, too, that in all three kinds of territory the system is also intended to assist the engineer in conforming to local restrictions, whether permanent or temporary. In dark territory there is the additional problem of hand switches; haven't tackled that one yet.

As for deciding what to do... in lit or CTC territory, there are quite specific rules as to what the various signals mean. Unless the signals themselves are poorly positioned (it has happened), I would doubt that it is possible to slide past a stop signal without breaking a rule, even in a heavy freight, unless there is an actual brake failure. (it used to be possible, before time and occupancy locking was universally applied) In any territory, however, it would be prudent to come up with rather conservative limits as to stopping capability, etc., and use them -- although one could get fancy and use the knowledge of the railroad built into the system and information about the train, obtainable from the existing data recording devices, to modify the routines.

[dharmon]

I'm not sure what you mean when you say a GPS beacon...GPS is passive. Whether it is integrated into a more complex navigational system or the little Garmin you use for hunting or fishing, GPS in and of itself is only a fixing source. Its accuracy is based on the number of satellites in reception, refresh rate and stability of the receiver. A moving reciever has less accuracy than a fixed one. The info from each sat signal is basically a line of bearing from the sat and universal time. All additional info provided by a receiver, location, elevation, course, speed track, etc.....are all derived from the fixes achieved from these lines of bearing, not the satellite signal itself and are a function of the software in the reciever itself. The GPS system relies entirely upon the sat signals in a broadcast manner. If it were two way, the number of users would drag the system down to the point that it was unusable.

Some sort of tranceiver to send two way voice, data or both is required or the GPS info is only of interest to the crew in the cab. The GPS system in place now, becasue it is receive only, does not provide any of this. A separate cell or sat data system is required to link the units and share data regardless of the input...GPS, track sensor, RFID

[dehusman]

You are making this way more complicated than it needs to be.

The locomotive has a two way GPS beacon on it. It tells the system who it is (what engine) and how fast its going. The signal system knows where all the signals are, whether they are cleared or not and whether the switches are lined against the train or not.

When the train approaches a signal, the signal system decides if it would be a restrictive signal. The signal system sends a signal to the ATS system on the engine and tells it the train has bassed a restrictive signal. If the engineer doesn't take action then the system creates a penalty application.

The hard part is deciding what to do to prevent a collision. How do you determine if a train is going to go by a signal? A passenger train can reduce to 30 and come right up on the signal before making another reduction and stopping. A freight train can be continually slowing and still slide by a signal. Throw in the train make up and the grade and its pretty difficult.

I doubt they'll be using RFID chips (that's what AEI is anyway, so every car and locomotive already has them on the equipment) since the that's the whole point behind the GPS system, two way communication. I can't see the RR's spending $100K at every CP to install AEI readers.

Dave H.

[kenneo]

We have a good thing going here! However, as Jamie mentioned, we need to keep on track, err, thread.

CTC has, hardwired, almost all of what we ave been discussing here. What it is lacking is a cost effective ATS.

The rest of what we have been talking about would apply to non-CTC territory be it dark or lit.

I wonder if any "suits" are reading this thread?

It would certainly be interesting to hear what the CSX Dispatcher has to say about what we are discussing here, as well as Houston Ed, Mudchicken and our several engineers and conductors that contribute. Folks????

[dharmon]

Please don't take my post as implying that we take the man out of the loop...far from it. Just as in aircraft, a person needs to be there to handle the unexpected....and in the simplest terms to look out the window. Radios and voice communcations either. The technology exists and has been in use for some time to issue orders and pass telemetry data to and from units in the field so to speak electronically. And onboard systems can be updated over the air to provide greater situational awareness to the crews. In fact, a screen can be set in the cab that not only graphically displays the crews route clearance, but can show them what's ahead and what their orders are simply by rolling a cursor over the unit ID tag. Now a crew can see what the opposing train's route clearance is also. A system can be as complex or simple as required or desired. Now getting sateliite bandwidth is a seperate issue.

[jchnhtfd]

Tree's comments remind me a bit of an old saw in the airline industry about really advanced aircraft control -- there will be two crew members, a pilot and a dog: the pilot's job is to monitor the computers, and the dog's job is to bite the pilot if he touches anything.

I quite agree with Tree, though -- it will be a long time, if ever, that ANY kind of automatic system or semi-automatic system can -- or, in my view, should -- replace the crew on the engine. That's not the idea. Nor is the idea to supplant the present CTC and centralised dispatching systems -- however much they may make one fume while sitting on a siding for nothing to happen, they are essential. The idea is to come up with an inexpensive, feasible system, using current technology, to improve safety by operating like an expanded ATS.

I don't think, incidentally, that the amount of information suggested in various posts above to be passed from the train to a central system and back would overwhelm a voice radio circuit; even in my first option a couple of posts above, all necessary information can be passed in about 20 'words' (computerese) each way, and should take less than a tenth of a second, max.

The key to my proposals above is the use of cheap passive technology to tell the train -- and hence the central system -- which track it is on; something GPS simply can't do.

Keep at it, guys -- maybe we'll come up with something somebody in management will have a go at!

[tree68]

The trucking industry is using two-way technology now, together with GPS. I know a mechanic who had a truck dispatcher a thousand miles away ask who had started his truck - they had to hot wire it to move it. Never mind using the train radio - it would get overwhelmed with all the data. The technology exists to move data via satellite to do everything that has been discussed.

It will be many moons before the technology exists to make the crew on a train redundant (there's a whole thread on that on the Rail Guide forum), and it will always be necessary for someone to intervene for the really close stuff (try making that hook from 1000 miles away, not to mention hooking up hoses), and for the unexpected, like kids on the tracks, but everything that's been mentioned can be applied to making the system more efficient and safer. They are, after all, just more tools to let people do their job.

[dharmon]

Unfortunately integrating GPs into any system doesn't solve the issue, because it is only one way information to the receiving unit. If you think of the GPS system (26 satellites worldwide I believe) as radio beacons, the receiver is taking this broadcats data from two or more and fixing it's position periodically depending on its refresh rate. The Area of probabilty or error increases with speed of the receiver so in reality the position is time late when it's displayed, although it may only be a second old. Becasue of the built in error on civil units, this doesn't give exact enough geo-loc to ensure that a switch is clear, which requires and addtional sensor to ensure both ends of the train are clear....which really doesn't alleviate the problem. There still has to be something to indicate this be it a transponder or tag scan like was mentioned above.

Similiar to some of the other posts, what you can do is create a railroad command and control type system, integrating an up/down link via sateillite or cellular tower and a GPS receiver. Central control can issue a train order and align switches and routes withour ever speaking to the train crew. A clearance can be issued electronically to the cab by central......sort of like an ATC aircraft route clearance, giving limits, restrcitons and such, but to a screen, similar to an aircraft flight management system. Route updates from central (washouts, wrecks, weather...any data important can be forward to the crew via electronic message. The train data ...fuel remaining, maint status can be transmitted by the crew also so the receiving end can prepare for servicing....airlines do this to turn planes at major airports.

The central computer would periodically poll all stations for location. The response would be transmitted from the cab to central (using train name or loco #) as ID tag and give the GPS derived location and speed. The central control computer or operator can align the route and deconflict. One problem with data link arrangements is that they are bandwidth intensive and the greater the number of users and higher refresh rate (polling) the slower it tends to run. It also is dependent on that central node staying up.

Sidings or any closely spaced tracks are still going to require some sort of track occpancy sensor that must be integrated into the system to allow or disallow the route clearance, which adds another level of complication, but then again, using an ATC model...yards are like local traffic patterns and controlled locally by a tower, and cross country routes are picked up by central once they leave the yard.

To "fail safe" it, so to speak..kind of like ATS if I understand it right, the crew would be required to acknowledge receipt of messages and if not, or exceeded the clearance limit, central could order the train to stop, bypassing the crew.

The military and availtion world use these kind of systems already. It can't be too hard to adapt them. Pricey but doable. But does it get you more than you have now?

[jchnhtfd]

Eric -- response to last question first: believe it or not, I'm currently employed as a computer programmer in data management... how I got there from being a Civil Engineer for a railroad I haven't a clue, but I did!

Many of the western lines -- including both CP and CN -- did have assorted oops detectors for slides and washouts. Trouble is, they were always where such things were expected to happen. Years ago, we put a train on the ground because a beaver dam gave way in northern Ontario; be nice to protect against those, too, but...

The beauty of using RFICs is that they are entirely passive (no battery, just a little chip!), and really really cheap. I quite agree that a combination of the two (RFICs plus GPS) appears to offer some real advantages. I suggest multiple RFICs on a track not for position information, but redundancy; they are so cheap that it would have a good cost-benefit ratio.

As I envision it, the lead engine would have the interrogator on board, and the GPS. The two, plus the speedometer, would be read by the lead engine's computer and would tell it where it was, which way it was going, which track it was on, and how fast it was going. One could go two ways from there.

Ther first would be to transmit the information (it isn't all that much, so shouldn't overload the train radio system) to a central dispatching computer, which would have all the 'knowledge' as to where every train was as well as all the various limits, temporary restrictions, and what have you as well as the current dispatching clearances, and that central computer would then transmit back to the train the appropriate speed limit for right now and the appropriate signal indication; the train's computer would take that info and display it and, if the engineer and conductor didn't follow up, apply a penalty braking. The biggest advantage to this approach is that there would be only the one data base to update. The biggest disadvantage (and it's a biggie: trust me, I work on these things) is that if the central computer goes down the whole system goes dark.

The other approach would be to have all the system knowledge reside locally on the engine's computer. You would still transmit the information to a central computer, but the primary purpose of the central computer would be to keep track of clearance limits and the locations of all trains, which it would simply retransmit to the various trains. The local computer would apply the logic to prevent conflicts. It might be possible to have the trains 'eavesdrop' on each other, and acquire the location information directly from each other. Assuming that could be done, and I think it could, the central computer would be primarily to keep track of clearance limits (which it does anyway in a modern CTC system)..The biggest advantage here is that it when the central computer goes down (which is less likely, since it isn't doing much), the system could be set to 'coast' -- that is, run with the current clearance limits -- until it could be brought back up. As you note, the biggest disadvantage would be to ensure that all engines had the current knowledge. For this I would be inclined to use a 1 gigabyte or so flash card (as is used in digital cameras): it would be the conductor's responsibility to ensure that he or she had the flash card reprogrammed in the yard office at the start of each trip; no big deal (you dould even set it up so that the train couldn't move unless the flash card was current). It would be no problem to put all of North America on such a card.

For 'happening now' events -- slides, washouts, derailments, who knows what -- the central system could easily transmit this in the form of an amended clearance limit.

I like this whole concept. I'll keep thinking about it!

[kenneo]

RFID -- Radio Frequency Identification Chips aka ACI tabs. When I was last working for a railroad, we were starting to put them on the cars, but the shortline was never going to put interrogators in. I had thought of this method, but there seems to be a bias twoards anything hardwired, as part of the RFID usually must be. I would think that an RFID system would work far better than a GPS system, but with a vast number of parts. And that may be the real reason it is not being implimented. With an RFID system, you have the same mainteinence headaches you would have with the current ATS systems - just a different format.

I saw a "bullet-proof" installation being tested once where the interrogator was imbeded in a tie and a control box with battery was burried in the roadbed at tie end. It had a radio transmitter that had its antenna up at the top of an old power pole that had been donated by the power company (I think - it was not new!). Also at the top of the pole was a steel box that protected a solar aray to charge the battery. I think that this installation was a lot of overkill, but I mention it to illustrate the point. They were testing components, I was told, and later the entire installation was removed.

Where would the unit transmit to? It could go to a central location via satallite, or it could go directly to passing trains, to both, or even to a signal(s). If you had a puzzel track installation, you would probably want to hardwire it. It would be a much simpler and more reliable installation then if radio was used. If you had the installation in CTC territory, you would probably want both the central location and the train to have the signal.

But there is a complication here that is not technical. It is manual. It is keeping (at least in the lead unit) the control computer on the locomotives properly updated. With run through power, this could be a real pickel. A failure here will cause a total system failure.

However, I think that perhaps the best arrangement here would be a combination of GPS and RFIC with the RFIC chips at the control points and the GPS doing the work inbetween. I would think the programming would be a lot simpler, and simple does equal fewer failures and these are then easier to repair. That way, a location that might have multiple event reporting devices (switch point position, derail position, track occupancy, signal displayed, etc.) could report to one transmitter that has a logic circuit attached that would give a "yes (green)", "perhaps (yellow)" or "no (red)" to the train at the proper location.

There are methods to protect against slides and washouts. The Hill Lines (GN, NP, SPS - I don't know about the Q family), the UP, the WP and the SP all had detectors for "land tilt and/or displacement", water flow detectors, limit switches for "bridge feet", tunnel and snow shed movement detectors, plus other kinds of detectors, so unless something should happen in a location that its occurance could not be forseen, protection would (could) be provided in this system also. These detectors were always connected to signals so a problem would always stop the train. Well, OK, if the engineer applied the brakes the train would stop.

Jamie, what think ye, lad?

BTW, what do you do to pay the bills? (OK, OK, $CDN, I know! Your work?)

[jchnhtfd]

It's Monday, and my brain is obviously not up to speed yet.

The solution to the 'which track am I on' question is relatively simple: there now exists a class of passive chip often referred to as 'RFID' chips. They are on a lot of freight cars and engines. Never mind general merchandise. They send a coded signal when interrogated by a transmitter, which gives information about the device. They are unbelievably cheap, (sorry Eric -- money always shows up somewhere!) and can be hardened to take almost anything.

So... what one does is place these on the track centre line at intervals of a few hundred feet to a thousand feet or so -- varying as trackwork complexity varies, but always more than one -- and these have a code which is unique to the track (and block?). The interrogator transmitter is on all equipped engines.

Now you have most of the ingredients: where am I? How fast am I going? What track am I on?

The last ingredient is: what is my clearance? In CTC territory this isn't really a problem, although it requires a downlink from the CTC system to the engine (which is well within current technology). It also requires a status uplink from the train to the CTC (am I in the clear?) which is also not a real problem. It's doable -- in fact, if any guys reading this thread are in operations on some outfit, I'd love to correspond about this!

There are some interesting questions regarding various system failures. Ideally, whatever one comes up with should fail operational; at worst, a system failure of some sort should be no worse than the old 'if you can't read the signal, you assume it is at its most restrictive indication' rule.

It does not, however, solve the problem in territory which is not fully centrally controlled: one of the worst wrecks in some years on CN was to a VIA train which was maliciously diverted onto a siding by an enterprising local youth at a hand-operated switch. This is, I think, manageable with present technology, but requires that every such switch, derail, what have you, have a nearly bullet-proof (literally, unhappily!), short range, self-powered transmitter which indicates the status of the switch (and would have to be fail safe -- that is, no signal or degraded signal equals worst case situation) and an accompanying receiver on the engines. Simple enough, but much more of a maintenance headache.

As I say, I'd love to talk about this with a railroad operating department, if anyone out there reads this who is in such a position...

Of course, none of this helps the poor wretch who's highballing along and comes upon some soft track or a washout where the rails are intact but the roadbed isn't...

[jchnhtfd]

The trick, of course, being to control an offending movement without impeding other movements.

Leaving out a few oddball situations, there are, fundamentally, three situations where an ATS type system is needed: first, excessive speed for the territory, such as a restriction on a curve or entering a downgrade. Second, excessive speed for a territory and condition, such as a switch lined for a siding in otherwise high speed territory. Third, and the one which I think is particularly in mind here (but they should all be addressed, if any of them are to be addressed), approaching the limit of clearance -- that is passing a caution stop at next or a mandatory stop signal or location.

The first condition could be handled fairly easily with existing hardware and software: the GPS tells the engine where it is, and software in the engine checks to see what the restrictions are in the territory (updated frequently, such as for maintenance) and checks its own speedometer and first warns the engineer and conductor and then applies braking if required. Note that the GPS is not relied on for speed -- the engine's speedometer is far more accurate! The GPS also informs on direction of travel, so that the speed can be controlled appropriately for upgrade/downgrade directions. While it isn't always true (is anything?), most commonly speed restrictions of this sort apply to multiple tracks on the same alignment equally, so determining which track you are on isn't a problem.

Both the second and third situations require knowledge of which track one is on. While I agree with Eric that most new track is on centres which may be wide enough to allow discrimination by a GPS, there is an awful lot of track out there which isn't -- and some of the most complex flavours are in metropolitan areas, where GPS accuracy degrades anyway. So the 'which track am I on' and 'which track is he on' issue is quite real.

It's Monday, and the office is real quiet. I'll think about this some more...

[kenneo]

Dharmon and Jchnhtfd

For discussion purposes, lets keep cost as a non-issue other than to attempt to provide whatever it is we are talking about at the lowest reasonable cost. What I was trying to do was have something that would get a train stopped prior to an Absolute Signal such as the Absolute Signals at Carrolls where they had that head on 8 years ago and another side collision earlier this month. The only differences were the crossover track to be used and the train that was to crossover. If we had some sort of "poormans" ATS that would have stopped the offending movement, in one case there would have been no sideswipe, and in the other, 5 men would still be alive. I guess what I am trying to do is to get the hot rodder stopped before he goes through the stop signal.

I have an idea, but don't have the knowledge of how it works today to be able to make any solid statements. The idea is to have "some sort of sensor arrangement" at each critical part of a control point group - whether it be an X in multiple track main or a simple siding of junction in single track. This would, at its most spare, be "Is the signal red?" then when the GPS on the train read a certain location from the CP, the approaching train would get a mandatory service application such as the old style hard-wired ATS does now, but would get it via wireless. The engineer would have to able to override because if we are dealing with ABS, the block signals are byond the points with the points pointed toward the signal rather than at the foul points as in CTC. --------:--\=================/--:---------------- -----.-\-:===== If ABS (1st diagram), and meeting a train on the main where the approaching train is to go into the hole, the approaching train would have to make a mandatory complete stop prior to the signals (the " : ") which is prior to the switch (the "\" ) even if the switch is lined for the siding, and then flag through the signal into the siding against the train holding the main. The approaching train would need to be in a position to see the signal at the switch where he was to take siding be red, then see it change to yellow as the headman on the train holding the main lined the gate for the hole.

One thought about the figure of 15 feet we have been using. When I was assigned to lay out alaignments for the shortline I worked for, we had to use a minimum of 22 foot track centers, which in theory, should put our transmitters with 15 feet of each other and no closer than 11 feet for a linear distance of about 10 or 15 feet. I guess what I am trying to say here is that the 15 foot parameter we have been using, in theory, shouldn't matter. But then trains shouldn't run into each other, either.

I guess the simple way to restate my question is, with off-the-shelf items available today, how can a system be put together (physical items and operating rules) that should be afordable so that these really stupid wrecks don't happen.

[dharmon]

And that becomes the rub so to speak or actually the prevention of it. With the current restrictions on accuracy afforded to civil users, the issue becomes that accuracy factor. A 15 foot Area of Probability or error is not sufficient to guarantee safety. Perhaps a transponder on the loco and FRED that activates signal, which in conjunction with a GPS position can confirm the ends of the train is clear of the switch. Which goes bck to a system which doesn't require a GPS. It can be done, but effectively what you are asking for is a sort of military style tactical data link which give position based on own unit reporting correlated by other units sensors as well as enemy units......this would require a fairly sophisticated (read expensive) integration of GPS data and track sensors.

[jchnhtfd]

Hmmmmm...[8)]. Well now. This is mostly to give the ball another gentle kick, to see what some of the others will say[:D]. There really wouldn't be a problem with single track operations. Clearly, if one is on a single track... and there is another train stopped ahead... or coming my way[xx(]... GPS could be used (presumably with an uplink to a central unit -- and what the heck, that's already used for general location and maintenance) and a downlink for commands to warn of the situation. Or, if a train is approaching a switch, it shouldn't be hard to assemble the info about speed and switch position to provide a control signal.

Where it gets interesting is on multiple track, or passing sidings, or junctions, where it is perfectly legal for two trains to be in the same place plus or minus the 8 feet or so -- provided they are on different tracks, or clear of the fouling point[8D].

My thought: suppose it were possible to uplink the GPS position and speed, plus the track identification (no, I haven't figured this out yet!) to a central logic unit somewhere (haven't figured that out yet, either). If it were, it would be no particular problem to send back appropriate control signals to the engineer and conductor (for instance, supposing you are on the main, westbound: you are about to meet Extra 8444 east, he's on East Overshoe siding, in the clear, west end switch lined for the main, you have a clear board. Or... Extra 8444 is moving into East Overshoe siding, not in the clear, you have a mandatory stop at mp 123.4. Or... Extra 8444 is holding the main at East Overshoe, west end switch lined to enter the siding, maximum speed 10 mph) (or, Extra 8444 is holding the main at East Overshoe, west end switch lined for the main, STOP NOW) or whatever.

The key to it is adding the information about what track the train is on, and any needed switch positions.

It could be done. Money. Time. Compatibility.

Anybody?

[kenneo]

Earlier conversations on these Forums have discussed GPS train separation. The main problem is the inability to determine location with an error factor of less than 7.5 feet Plus or Minus (15 feet total).

Since the current ATS is expensive and always in need of repair, within the limitations of GPS technology today, how could a "GPS ATS" be put together?[?]