Login
or
Register
Home
»
Trains Magazine
»
Forums
»
General Discussion
»
Simple
Edit post
Edit your reply below.
Post Body
Enter your post below.
[quote user="jeaton"] <P>Use of GPS to pinpoint the location of trains has sort of become the holy grail for some who think it must be part of the ultimate crash avoidance system. Knowing where a train is located is only a fairly small part of the process of getting trains moved over the railroad. Even if GPS could be used to identify the exact location and there were no signal loss problems, it is still providing no more information than that which is known by the train crew. The big part is moving the location information from the train to the dispatcher and the electronic support to his decision making process, then getting instructions back to the train crew and finally getting a feedback to the dispatcher indicating compliance with the instructions.</P> <P>There are many systems in place that accomplish the goal of keeping trains from running into each other, and those involving GPS are not necessarily the better mouse trap. It has been authoratively reported that the train control system designed for installation on the formerly dark Iraq Railroad had a cost of about 20% of "conventional" systems with lineside signals and it did not involve GPS.</P> <P>Cost and usefulness! You may notice that railroads have also taken a pass on another recent develpment and haven't put flat screen HDTV set in locomotive engine cabs. </P> <P>[/quote]</P> <P>I realize that train-control is an arcane science even inside railroading, and much of what has been distributed to the public is badly written if not incomprehensible, so please don't take offense if I point out that GPS is already in use on U.S. railroads for train-control applications and it does have value, if not silver-bullet value.</P> <P>The problem with GPS is that it provides physical values whereas train-control requires logical values. This is an extremely important distinction that elides most observers of railroads. Because trains are constrained to tracks (a logical system) train-control requires logical inputs. GPS provides a physical input that says "you're sort of here, most of the time," whereas a train-control system says "you're either here or not here, all of the time." Because there's no way to measure physical location precisely or accurately, electronic train-control systems use logical either/or methods, such as track circuits (either open or shunted) or transponders or inductors (either passed or not passed). Manual train control systems such as train orders or TWC use manual logical inputs, e.g., "I am on the main track," or "I have left the main track." Yes, electronic circuits can fail but there are fail-safes built in plus redundant paths that makes shunted (occupied) the default condition and open (not occupied) the special condition. "Failure to shunt" resulting in a collision or injury is extremely rare.</P> <P>The issue with GPS is that it's fuzzy, and while most of the time it will discriminate which track a train is on, and which side of a track cut (insulated joint) a train is on, it cannot do this reliably. (Do you want to get on an aircraft with a 1 in 10,000 chance of crashing or a 1 in 1,000,000 chance of crashing?)There are ways to get around this. A transponder tag on the track will tell the system which exit a train has taken from a control point. The system architecture knows there's a tag there and if the GPS indicates the train has run 150 feet past the tag location and the train hasn't not found it (because the tag has vanished, usually), the GPS can no longer "map" the train to a specific track, the system goes to a failed state, and the train is braked to a stop as well as any movements that may be opposing it. The way around determining which side of a track cut a train is on is to only care about one side of the track cut. You set one side as controlled track and the other side as uncontrolled track, and set a GPS fence where any train approaching the track cut is assumed to be continuing to the other side of the track cut, and predictively brake any train to restricted speed as it approaches the track cut heading toward uncontrolled territory, and predictively brake any train to a stop as it approaches the track cut heading for controlled territory. As is obvious these solutions to GPS fuzziness greatly reduce track capacity and train speeds, and neither will absolutely prevent collisions between trains that violate their authorities, just make them occur at low speeds.</P> <P>GPS is used in CBTC (communications-based train-control) systems, which is an outgrowth of the BNSF Generation 1 ETMS (electronic-train management system), an outgrowth of the CSX CBTM (communications-based train-management system), which came from the Rockwell-Collins ARES system tested on BN about 15 years ago, which was the first system to employ GPS as a method of eliminating track circuits for train-location detection. (ARES worked very well, by the way.) Rockwell-Collins spun off their train-control effort to Wabco which later renamed itself Wabtec. ETMS uses differential GPS to determine train location, and RF interfaces with wayside signals, switch-point indicators, and electric locks to determine the condition of wayside devices. Peer-to-peer communication is used to reduce transmission and processor loads. ETMS is pending FRA approval <EM>only as an overlay system to traditional methods of operation such as CTC and TWC, not as a stand-alone method of operation. </EM>The ETMS overlay system provides an order-of-magnitude increase in safety by providing predictive braking of trains approaching authority limits and permanent or temporary speed restrictions. FRA approval of ETMS or CBTC as a method of operation hasn't even been asked for yet by any U.S. railroad. Federal law (49 CFR Part 236 Subpart H) requires a railroad proposing a new method of operation (e.g., CTC, DTC, TWC) to prove that the new method is as safe or more safe than the existing, plus propose in detail how the railroad intends to manage the implementation and maintenance of the system. That proof is both expensive and time-consuming; the <EM>railroads are not at fault for this. </EM>One could say the FRA is not very fast on this, but Congress and the White House have given them a law and a personnel incentive system that rewards ultraconservativism<EM>.</EM> But frankly, management of methods of operation is vastly more expensive and complicated then even most people in the railroad industry realize, and making changes in them is a high-risk, high-cost undertaking.</P> <P>The Iraqi system, as I understand it, is a partial implementation of the BNSF ETMS system. It also uses GPS to map trains onto a GPS database of the track layout, plus transponders to identify control-point exits, but unlike ETMS it gives up <EM>all </EM>control of turnouts and sidings. That is, the Iraqi system only controls main track between stations (or main track between sidings, in U.S. practice). Stations are treated as non-controlled track operated only at restricted speed (yard limits, in other words) and trains are required by rule to brake to restricted speed before entering a station. The Iraqi CBTC system reactively brakes trains that do not reduce speed appropriately upon approach to a station (reactive instead of predictive because train consist and braking equipment condition is not known). But an improvement that ETMS doesn't have is that instead of merely displaying the authority to the train granted by TWC or CTC, it bypasses that step and <EM>takes</EM> the authority from the dispatching computer, and enforces it. In other words, think of a TWC or DTC form that once generated by the dispatching computer, instead of being read out loud over the radio and copied by pen, it was digitally transmitted to the locomotive to appear on a screen, and the engineman pressing an "accept" button replaced the readback. Several electronic "handshakes" later and you have a valid authority. (ETMS doesn't have this because of Subpart H.)</P> <P>There's no reason you couldn't have a GPS-based train-control system in the U.S. as a stand-alone method of operation, except, of course, is that you'd have to meet Subpart H and you'd have to give up capacity compared to traditional CTC, even if you put switch-point indicators and RF interfaces on all the turnouts. You would also have to heavily rewrite GCOR or whatever rule book you were operating under and retrain everyone governed by the rule book, and that's an onerous and risky undertaking. But you would have an enforced speed-limit and authority-limit system, and right now that's something TWC, DTC, and CTC won't do (unless they have an ETMS overlay). It also would mean that "missed repeats" would be a thing of the past. It has a lot of attractiveness for low- and medium-density dark territory but not so much for high-density, as a stand-alone system.</P> <P> S. Hadid</P>
Tags (Optional)
Tags are keywords that get attached to your post. They are used to categorize your submission and make it easier to search for. To add tags to your post type a tag into the box below and click the "Add Tag" button.
Add Tag
Update Reply
Join our Community!
Our community is
FREE
to join. To participate you must either login or register for an account.
Login »
Register »
Search the Community
Newsletter Sign-Up
By signing up you may also receive occasional reader surveys and special offers from Trains magazine.Please view our
privacy policy
More great sites from Kalmbach Media
Terms Of Use
|
Privacy Policy
|
Copyright Policy