Why signals? Why not GPS + radio network?

GOLDEN-ARM wrote:Just curious, Russ, you don't happen to sell, maintain or install the GPS type systems, in question, do you?

What would be different if I did? Would you treat a railfan differently if you knew he worked for a competing railroad?

But regardless, you've given me a reason why signals are better than a GPS+network: because the signalling system can detect wayward equipment.

RussNelson wrote:[...you've given me a reason why signals are better than a GPS+network: because the signalling system can detect wayward equipment.

...and broken rails and switches lined improperly.

No, it wouldn't matter if you were a railfan for a competing railroad. (those of us in the field don't really give a damn, about the competition. a train is a train is a train...) It would make a difference, however, if you were interested in promoting ANYthing, that was "self serving", so to speak. If you were a salesman, of GPS parts, or related services, I would be a bit more guarded, in conversing with you, if I suspected a "hidden-agenda", in a conversation. Even though you didn't answer my question, I have still answered yours. Regards

GOLDEN-ARM wrote:Even though you didn't answer my question, I have still answered yours. Regards

Do you have that much of a problem with vendors promoting railroad products on railroad.net? It looked to me like you were trying to dismiss my suggestion as one coming from an interested party. I"m interested, sure, interested in promoting railroads. I've heard (but nobody suggested in this thread) that you need signals to run passenger trains. Seems to me (but I was wrong) that GPS+network would be the equivalent to signals, so that could bring back limited passenger trains to dark territory, where appropriate. E.g. on the Montreal Secondary (UPDATE: apparently now called the St. Lawrence Subdivision) there's only four trains a day. What if you could run a train between Potsdam and Canton? Just back and forth, all day long. It would be more energy-efficient than a bus, and the train station in Canton is closer to campus than the bus station.

But anyway, iit won't work, as you pointed out.

RR.Net is not "mine", to guard. That being said, I am sure vendors are welcomed, with open arms, in the appropriate venue. The header on top of this very thread, is there for vendors to display their wares, for all to see. The site is run on money, and a person profiting from the sale of items, goods or services would more than likely have to compensate the owners, of THIS site, for the priveledge of advertising here.
I would not imagine why a train could not operate in dark territory, with a track warrant/train order, specifying limits, and directions. The real reason buses now carry passengers, in places where trains once ruled, is a simple economic decision, made by the carriers. Trains were/are too expensive to operate, in some areas. They thrive, where there is sufficient customer base, for the service, and they languished, and finally died, where that base was insufficient. It's hard to justify a quarterly operating loss, in the thousands of dollars, so a few folks can ride a train. I would like to see a return to the passenger trains, and the commuter trains of earlier decades, but that's probably a thing that will not come to be, now or in the foreseeable future........

The largest obstacle to passenger service is the capital required for equipment and track structure.

First, I don't think I'd put too much stock in any tech company trying to advance their cause on RAILROAD.NET - with all due respect to this website... that said:

A "GPS unit" is a bit of a misnomer. This would actually be telemetry technology that has incorporated into the hardware a GPS unit. The GPS unit merely "talks" with government satellites to determine location. Those satellites aren't going to tell you anything but lat-lon coordinates. Even if you have a database onboard a locomotive to interpret the lat-lon into a location, e.g., milepost, SPLC, etc. (called reverse-geocoding), you still need to know where the other trains are. Enter the transceiver that communicates with a central server through either satellite or terrestrial (cellular) means.

Telemetry units are already installed on locomotives (to monitor performance) and reefers (to monitor reefer unit performance and inside temp). They are (slowly) finding their way into the general freight car fleet in part because railroad car location messaging is so slow and inaccurate.

When/if telemetry technology makes its way into dispatching systems, who knows... sure, there are many technical issues to surmount, but it's naive to think that this technology could not be incorporated into a "failsafe" system. Heck, the current technology relies on copper wire (that gets stolen) and electricity- no-one reading this can claim they've never suffered through a power outage.

Take a look: Positive Train Control

arinc.com wrote: Benefits:
* Enhanced operational safety and line haul capacity
* Improved operational efficiency of track forces and dispatching
* Improved locomotive performance and operating costs
* Reduced-cost system expansions

GOLDEN-ARM wrote:Those GPS systems don't detect broken rails. Signals can, and do.

I actually saw two rails joined with one of those bolted steel plate brackets, and I'm not kidding ya...that bracket was completly split apart...yet electrically the continuity was there cause signals were operating fine, and trains were going over it...... So I guess we don't have 100% fail safe in the track circuitry either do we?

DutchRailnut wrote:reliable is not enough, a signal system needs to be fail safe.
how about loss of signal from satellite ?? how about interference ???
not even one nano second can it be compromized.

That argument is fundamentally flawed! Humans are error prone so there goes your fail safe arguement. The Toronto Transit Commission had an under ground collision cause a subway train driver choose to drive on past two signals with red aspect....and as luck would have it the trip arms were broken. Get rid of the humans, bring in 100% automation and then we will be must closer to fail safe.

Wrong since computers are only as good as programmer, and automation as good as its components and the mechanic working on them you never elliminate the human aspect.
Your bolted rail may have had broken plate but was it broken on both sides of the rail, did it have a bonding wire welded to the rail. as long as there is continouity the signal system is intact

DutchRailnut wrote:did it have a bonding wire welded to the rail. as long as there is continouity the signal system is intact

Yes it did, and thank you for making my point. The signal system, more specifically the broken rail circuitry, is not fail-safe because we humans bonded a wire across a piece of rail. We didn't expect it to break there, but it did.

Nothing is fail-safe.

DutchRailnut: Actually, we have eliminated the human from many aspects of computers. Humans don't design chips anymore. Humans don't write assembly language anymore (except in certain specialized situations). Yes, a bug in a compiler is a nasty problem to encounter. But airplanes can take off, fly, and land completely under computer control. In some ways the rail network is a more predictable environment than air travel. In other ways, it's more predictable in a bad way -- ask any engineer who can see an auto stopped in a crossing a mile away. He knows he can't avoid it.

As LCJ pointed out, my original premise was flawed, however, there is good reason to use GPS + a data radio network as an adjunct to a signalling system.

Some interesting thoughts here.......

My issue isn't who RussNelson works for or if he's trying to sell products to railroads. My issue is one that has been echoed here already-safety.

RussNelson wrote:

Yes, it needs to fail safe. Thus, every train and dispatcher knows where every other train is. If a train loses connectivity with the network, it will run according to a pre-determined schedule. Thus, every other train will know where that train MUST be. (MUST be?-then reality set in...)

Yes, there will be GPS satellite outages, but they will be predictable based on location. If the signal goes away when the train enters that location, nobody gets upset. The position estimation algorithm will assign an error value, and when a train gets too close to a silent train, it will stay far enough away for whatever safety margin you want. The real problem is not when one node goes off the air; it is when your entire network fails, and every train disappears, and you lose voice contact with the trains at the same time.

There are far too many variables here......I worked for about a year as a train dispatcher trainee for a large northeast commuter railroad. We controlled our trains via CTC. If the signal system went down we operated by written train orders. Written train orders keep trains safely separated by only allowing one train in a block (or blocks) at a time. Each train has to stop at the end of his authorized clearance before he will receive an order to proceed to the next block(s). In addition, he will only receive that clearance if the block(s) ahead are clear. It is time consuming, it does slow operations down, but it also seems to be a much more positive way of keeping track of where your trains are when your primary dispatch system crashes.

The system RussNelson describes sounds good but there are too many variables to take into account if the system were to crash. CTC backed up by written train orders seem to be the safest way to control trains, especially those that are carrying hundreds of commuters at a time.

Just my two cents!

LCJ wrote:...and broken rails and switches lined improperly.

Not entirely true. There was a bad derailment of a VIA train here in Nova Scotia because a switch was not lined properly. From the resulting investigation I believe it was concluded that a switch can be up to a 1/4 inch from complete aligment, and still show (visually and electrically) a completed alignment. Once you go past the 1/4" the points change thus changing the info sent to the signalling system.

In the VIA case the switch was within the 1/4", the CTC system showed a correctly lined switch, the train went into this section at its normal speed and then derailed. The train went through a building that had people in it, but thankfully no one was killed. Fail safe eh?

A teenager was found to have smashed the lock off, and had moved the switch slightly. I believe CN concluded that it would be too costly to make any changes system wide, and thus concluded the solution would be too use a higher 'caliber' of lock that cannot be so easily smashed off.

http://www.cbc.ca/story/news/national/2 ... 10414.html

Reference aircraft, the FAA intends to eliminate all radar and rely on WAAS GPS beacons from aircraft for aircraft position reporting to air traffic control. Radar is expensive to maintain and there are many areas that are "dark" in railroad speak. Note that most aircraft do not have on board radar so that is not a practicable alternative means of achieving separation.

That said, aircraft GPS does not suffer from poor GPS signals due to terrain, buildings and foliage. An aircraft, except if flying low in a mountain pass or canyon, has an unobstructed view of the GPS satellite constellation at all times. So it can be reliably and safely used. However, any ground based GPS vehicle encounters numerous instances of loss of GPS accuracy or entire signals which would preclude its use as a primary control system for a busy railroad, I would conjecture.