In the excellent article about the Quebec North Shore & Labrador in this month’s Trains, mention was made (on page 44) of the use of Proximity Detection Devices (PDD) that the Railroad uses. The article states that QNS&L is the only railroad to use them and that apparently the engineers love them.
Why aren’t other railroads using this technology? From what I’ve read Postive Train Control is supposed to prevent collisions and allow for more efficient operations; however, PTC is very expensive and involves technology that hasn’t been fully developed or proven yet. Wouldn’t PDD be a more cost effective alternative to PTC? I’m curious as to why PDD aren’t being used elsewhere.
Works OK for a single track low density operation where it serves as an additional alert for a 1 man operation. The engineer may have been running for hours in very boring circumstances.
The system does not have the precision that would be necessary in high density lines. Say there is 20-30 trains per day each way on double track. If the “train in proximity” alert required reduction from track speed, trains running in such a situation would be spending way to much time at medium or restricted speed just to pass a train approaching on the opposit track.
The “Positive” in PTC also means that the engineer cannot override a system order to reduce speed or stop.
Darn glad you raised this question, because I’ve been meaning to since I read that article last weekend . . .
The aspect (no pun intended) or scenario that I was wondering about is: How does the PDD know the difference between the rear - EOTD ? - or front of a train ahead on the same track, and the head or rear going past another train in a passing siding only 14 to 15 ft. away ?
However, on re-reading that portion of the article closely, I see that all the PDD does is beep, and require the engineers of both trains to acknowledge the beep - or a “penalty brake application” will occur - at distances of 8, 5, and then 3 miles. So by merely acknowledging the beeps at the proper time, a train can easily get by another train in a passing siding - or proceed right into the rear of the train ahead !
In that regard, the PDD is not too different from some ‘train stop’ systems - there’s no decreasing speed control as the trains approach closer to each other, and as long as the engineer acknowledges the system’s beep/ buzzer/ horn/ alerter/ whatever within the prescribed time, he can continue to proceed at full speed.
That said, it would seem that acknowledging a beep 3 times within every 2 - 3 miles would ensure sufficient alertness and situational awareness that the self-preservation instinct would kick into gear to prevent a collision at which that engineer would be “the first one on the scene”, so to speak. [:-,] But again, that’s about the same as a train stop system, and those are not fool-proof or immune to that type of human failure.
Still, the PDD is better than nothing. Otherwise, I believe this line is CTC-controlled, but that system grants and restricts track occupancy authority by signal indication only and the presumed obedience by the solo engineer. The CTC system h
From what I read in the article, when it issues a warning it dosen’t require reduced speed. Only to acknowledge the warning. If not acknowledged then it will make a brake application. It’s kind of like some cab signal and train stop systems. You need to acknowledge the changes in aspect, but it dosen’t require you to actually slow down or even stop. The idea is that a person would heed the warnings and operate accordingly. The thought that a sleepy engineer might acknowledge the system and fall back to sleep without reducing speed or preparing to stop isn’t figured in.
That’s why the “positive” systems will require more than just pushing a button or moving a lever. You’ll have to also run the train within the parameters the computer figures is safe for a given situation.
I didn’t mean to imply that the system is of no benefit in the circumstances of the operation reported in the article. It is a redudancy to the CTC and/or Track Warrant control and for safety sake redundancy is usually good. I suspect an “alert” triggers an exchange between the trains with clarification of status. In the event one of the engineers was disabled and couldn’t respond, the other train would proceed at restricted speed, prepared to stop as might be required.
Given that the railroad is essentially isolated from a national system and a relatively low requirement for hardware, I would think that the cost /benefit ratios are good. I just don’t see it providing that much added benefit to a busy mainline railroad with a two man engine crew. One can consider the second man in the locomotive for most US freight operations as the safety redundancy. Although that clearly does not produce absolute protection train collisions, it would seem to be pretty close.
An added thought. A train operating in a busy terminal area such as Chicago would probably be under a constant proximity alert set up by other trains not even on the same railroad. What would be the rule in those circumstancies? Proceed assuming all is well, or run at restricted speed? The former renders the proximity alert sytem of no benefit, the latter would certainly increase the cost of getting trains moved throught the area.
Jeff, thanks for confirming my understanding of the limitations of these systems. [tup]
In response to jeaton’s post above: Perhaps use different frequencies or digital ‘codes’ for different railroads to prevent excessive ‘false beeps’ ?
The though also occurs to me that much of what Jay wrote above about the PDD could also be - and has been said - about PTC: not much added benefit, increasing the cost (by decreasing speed and capacity), etc. [:-^]
We have had a system like this on airliners for 22 years. It gives us a 360 degree view that goes out 40 miles. ( 15 miles on Southwest ) It will show us any traffic that is within 8000 feet of our altitude. It then gives us prompts to avoid a collision, such as " Don’t Climb " or " Descend, Descend ". The prompts are both aural and red, yellow or green lights on our VSI, Vertical Speed Indicator. We too have problems with false warnings in the terminal areas where many aircraft are in close proximity near the airport. The aircraft system may be more sophisticated than the RR system in that the two boxes talk to each other so they don’t give conflicting prompts. IE, both aircraft prompted to " Climb ". Only TCAS II systems talk to each other. All aircraft appear superimposed on our digital radar scopes. There is a readout next to each target telling us the aircraft’s altitude and an arrow telling us if it is climbing or descending. It is interesting to note the TCAS, Traffic Collision Avoidance System, is compatible with the marine system and will show us ships when we are below 8000 feet.
Geing a progammer from way back (over 46 years). The PDD software could be refined to only require action for the track on which a train is running, not on adjacent tracks. This would it to be usefull on high densicty tracks. Commercial GPS is good for within a 3 foott radius of its target and paralell tracks are more than thaat.
That’s half of the challenge faced by the folks developing the new mandated PTC. The PDD as used by QNS&L really only needs to know there is another train somewhere within range, very simple.
In major cities especially, and anywhere there are multiple tracks, as well as knowing exactly which tracks the trains are on, it also has to know that there are no crossovers in the interval, or junctions, or diamonds instead of grade separations. That’s a whole new layer of intelligence required, orders of magnitude beyond what the QNS&L is using.
That’s half of the challenge faced by the folks developing the new mandated PTC. The PDD as used by QNS&L really only needs to know there is another train somewhere within range, very simple.
In major cities especially, and anywhere there are multiple tracks, as well as knowing exactly which tracks the trains are on, it also has to know that there are no crossovers in the interval, or junctions, or diamonds instead of grade separations. That’s a whole new layer of intelligence required, orders of magnitude beyond what the QNS&L is using.
Nope - not at track speeds (“Real-Time Kinematic” or “RTK” processing mode), and especially not with all the multipath possibilities of the signal bouncing off adjacent metal cars, almost any warehouse or industrial building, satellite ‘shadows’ (northern side of) of mountains and trees, tunnels, any deep cut, in through-truss bridges, under highway bridges, power lines, etc. It’s hard enough to get reliable readings for even rough topo survey work - i.e., 3 ft. accuracy - in some of those conditions, when standing still.
Perhaps the best way of determining train location is to have a bunch of RFID tags along the right of way that will give info on which track the train is on and an indication that the front (or rear) of the train was within a few feet of the tag. Tag density would be much higher near switches and crossings. GPS would be used to gauge progress between tags (along with some sort of dead reckoning - either wheel or doppler radar based) and call an alert when a tag is expected and no tag is seen.
Getting a consistent 3 foot accuracy with satellite nav would require a higher chipping rate on the signals, broader receive bandwidth, higher satellite transmit power and at least two separate frequencies. Doing that would require 10’s of billions of dollars for the new birds, plus whatever it takes to remove users from bands adjacent to the current GPS bands. Considering the brouhaha surrounding Light Square and GPS, the latter may be the biggest challenge. (FWIW, I’m with the GPS communty in the LSQ vs GPS battle.)
Note that there would be a lot of places where this “enhanced GPS” would not work for pretty muc the same places that Paul gave for the current GPS not working. One possible exception is that multipath can be reduced by using an electronically steerable antenna.
The GPS accurcay used by the Tomahawk, Minuteman and Trident misles is many times more precise than commercial GPS. How much I am not allowed to say as it is still classified, but the chips are available to allow for accuracy on multiple tracks. The idea of using RFID is good, if used with a combination of enhanced PDD and PTC to tell the locomotives who is near them and not allow them to pass a yellow (Approach) signal without reducing speed or a red signal under any circumstances without first stopping. I agree tht it would take enhanced programming into the cips, but it is not that hard with good programmers. A combination of Asembly language for speed and Fortran for the number cruching would do it nicely.
"The GPS accuracy used by the Tomahawk, Minuteman and Trident missiles is many times more precise than commercial GPS. How much I am not allowed to say as it is still classified, but the chips are available to allow for accuracy on multiple tracks. "
And until the defense establishment is willing to declassify the information it will be unavailable to the railroads. Once the chips are released for public commercial use, the rest of the world will also have access to the better system. The railroad environment does not have the tight security control that a missile base does.
The regular railroaders among us can help out, but I’m pretty sure that intermediate block signals with a number plate are permissive: A train may pass a red signal at a speed not to exceed 15 mph expecting to find another train or other obstruction in the block. Stopping distance must be one-half of sighting distance.
That’s what I was thinking about. I don’t know that all railroads have gone to that, some may still require a stop before proceeding at restricted speed. Depending on the railroad, the red signal with a number plate may now be called a “restricted proceed” or a “restricting.”
An old BN time table I have didn’t show a “stop and proceed.” The latest BNSF signal info that I’ve seen (about 4+ years old) had the “stop and proceed” again. Maybe they have changed back, no longer requiring a stop. There also could be local instructions in place that allow a train to pass without stopping, not to mention some conditions under the rules where a stop isn’t required where it otherwise would be.