 |
Re: GPS and PTC? Too much reading?
Author: OldPoleBurner
Date: 08-25-2011 - 11:57
I believe that in the past, Mr Meatball has identified himself as a lead radio communications engineer working with military satellites. That is his full time job! He must have intimate knowledge as to how satellite communications work and how they don't.
But whether or not he has that experience you speak of, I do! Several thousand hours of it over several years, in fact; and that was direct experience with systems trying to use the GPS, combined with ground based radio ranging, to accurately and failsafely locate and control trains.
The goal of the project I worked on, was to completely replace virtually all aspects of existing signaling and train control technology, from the track circuits all the way up to the onboard controls and the signals themselves, with a thoroughly modern CBTC system. But after several years of effort by several agencies, a checkered involvement by several private vendors, the FTA, and in excess of 150 million in (wasted?) Federal dollars; that effort was scrapped. Sadly, nothing was salvaged for that 150 million! Except maybe the experience and the knowledge of a whole lot of things that will not ultimately work - at least not in a safe enough manner.
I suppose that it is quite OK if some railfans, want to ignore all that collective experience. But in the real world of science and engineering, or even in the business world, you had better not ignore anyone's experience - not any - none.
Success in these endeavors comes only by doing less of what doesn't work and more of what does! You had better come to know which is which - fast. At length, I finally quit involvement with these sorts of projects; when I realized that most of those in the lead had lost their essential scientific objectivity; and were in fact, ignoring very fundamental problems - putting off solving them while continuing work in easier areas - merely accepting as a given without testing, that certain proposed methods would work somehow. This despite many warnings about flawed untested assumptions. Those issues turned out to be unsolvable, without resorting to the use of every single element of existing train control and signaling systems that ever existed - including picking up fixed track beacons and coded signals via the rails - and that was contrary to the project goal.
But the last straw for me was when those vendors refused to put their company liability on the line, specifically refusing to safety certify any application of any radio ranging technology for use in directly controlling trains. They were however, quite willing to certify it for non-safety critical uses. It also was the last straw for the FTA, who continued to insist upon safety certification. Other vendors were sought and they too ultimately bowed out - one after another after another.
Ultimately, it was a visit by the Doctorate Physicist, in who's name the patent was held for the underlying technology of GPS; that came and explained in person, why the private vendors involved refused to safety certify their products for use in safety critical applications. He told us bluntly, that in a military battlefield situation, losses are expected, and tolerated as inevitable; but that in the public transportation realm, there must be zero tolerance for safety critical errors, and that radio ranging technology inherently could not meet that standard. It can and does make errors routinely - some of which will be safety critical.
So what about PTS? Amtrak already has stretches of 'ITCS' in service - which could be described as the closest thing to GPS based PTS, that there is already in service. But this system only adds track beacons and GPS to the existing (but beefed up) coded track circuit systems. The trains must still pick up data from those track codes in order to fully function as a means of collision avoidance. Moreover, most of the system resources are devoted to continuous diagnostic and self testing algorithms; and when the P.U.F. (Position Uncertainty Factor) of the GPS location gets too great, or other system errors occur; it drops off line, resulting in a total "fall-back" to the conventional signaling system.
So while you can get that 'PUF' down to a only a few feet for a stationary object, if left in place long enough; it will not stay there for long while in motion. But in the practical world at 100mph, the 'SAFE' uncertainty factor ranges from +/-100ft at best to +/-350ft typical to infinite at worst; which happens anywhere there are metal objects, bridges, after passing under an overpass, or in tunnels (with ground based transponders). In fact, the critically important radio propagation rate inside a steel re-enforced tunnel or subway is totally unpredictable - one of those unsolvable circumstances (Yep, contrary to the junk science commonly taught in public schools, it is not a truly fixed constant - and neither is the speed of light, actually.)
The bottom line is, that while you can get GPS based PTS to work, you must allow several hundreds of feet extra margin for safety. You must also provide a fully functional coded audio frequency track circuit type signal system, as a basic foundation, in order to make it work at all, for failsafe train separation purposes. Moreover, the use the GPS data, is strictly limited to allowing the train to position itself when between wayside beacons. So it does not soley rely upon remembering the last wayside beacon - nor does it rely soley upon the GPS.
The big invisible elephant in the PTS room, and why I object; is that a specific vendor's technology has been mandated by political fiat, which has zero service record (not Amtrak's ITCS already in service, btw), The 2nd even bigger elephant is that conventional signal systems must still underlie the full use of PTS. However, this far cheaper conventional underlying system is in itself fully capable of all the safety sought after - without any so-called PTS at all.
So while it may be nice to have all the extra data in the cab that a GPS based system would afford, it is not essential to safety (the stated purpose of the mandate); so should not have been mandated. The free market could then decide if all the 'nice-to-haves' were worth the cost. What should have been mandated at a minimum, was the use of any system that failsafely and continuously supervised braking, while enforcing a not-optional stop prior to passing any stop signal.
Moreover, That mandate should have been applied based upon potential severity of ANY accident - in other words, maximum territorial speed - perhaps anything 30mph or over, unless permitted traffic in a whole territory was limited to only one train at a time.
But enough already - I'm still employed!
OPB
But whether or not he has that experience you speak of, I do! Several thousand hours of it over several years, in fact; and that was direct experience with systems trying to use the GPS, combined with ground based radio ranging, to accurately and failsafely locate and control trains.
The goal of the project I worked on, was to completely replace virtually all aspects of existing signaling and train control technology, from the track circuits all the way up to the onboard controls and the signals themselves, with a thoroughly modern CBTC system. But after several years of effort by several agencies, a checkered involvement by several private vendors, the FTA, and in excess of 150 million in (wasted?) Federal dollars; that effort was scrapped. Sadly, nothing was salvaged for that 150 million! Except maybe the experience and the knowledge of a whole lot of things that will not ultimately work - at least not in a safe enough manner.
I suppose that it is quite OK if some railfans, want to ignore all that collective experience. But in the real world of science and engineering, or even in the business world, you had better not ignore anyone's experience - not any - none.
Success in these endeavors comes only by doing less of what doesn't work and more of what does! You had better come to know which is which - fast. At length, I finally quit involvement with these sorts of projects; when I realized that most of those in the lead had lost their essential scientific objectivity; and were in fact, ignoring very fundamental problems - putting off solving them while continuing work in easier areas - merely accepting as a given without testing, that certain proposed methods would work somehow. This despite many warnings about flawed untested assumptions. Those issues turned out to be unsolvable, without resorting to the use of every single element of existing train control and signaling systems that ever existed - including picking up fixed track beacons and coded signals via the rails - and that was contrary to the project goal.
But the last straw for me was when those vendors refused to put their company liability on the line, specifically refusing to safety certify any application of any radio ranging technology for use in directly controlling trains. They were however, quite willing to certify it for non-safety critical uses. It also was the last straw for the FTA, who continued to insist upon safety certification. Other vendors were sought and they too ultimately bowed out - one after another after another.
Ultimately, it was a visit by the Doctorate Physicist, in who's name the patent was held for the underlying technology of GPS; that came and explained in person, why the private vendors involved refused to safety certify their products for use in safety critical applications. He told us bluntly, that in a military battlefield situation, losses are expected, and tolerated as inevitable; but that in the public transportation realm, there must be zero tolerance for safety critical errors, and that radio ranging technology inherently could not meet that standard. It can and does make errors routinely - some of which will be safety critical.
So what about PTS? Amtrak already has stretches of 'ITCS' in service - which could be described as the closest thing to GPS based PTS, that there is already in service. But this system only adds track beacons and GPS to the existing (but beefed up) coded track circuit systems. The trains must still pick up data from those track codes in order to fully function as a means of collision avoidance. Moreover, most of the system resources are devoted to continuous diagnostic and self testing algorithms; and when the P.U.F. (Position Uncertainty Factor) of the GPS location gets too great, or other system errors occur; it drops off line, resulting in a total "fall-back" to the conventional signaling system.
So while you can get that 'PUF' down to a only a few feet for a stationary object, if left in place long enough; it will not stay there for long while in motion. But in the practical world at 100mph, the 'SAFE' uncertainty factor ranges from +/-100ft at best to +/-350ft typical to infinite at worst; which happens anywhere there are metal objects, bridges, after passing under an overpass, or in tunnels (with ground based transponders). In fact, the critically important radio propagation rate inside a steel re-enforced tunnel or subway is totally unpredictable - one of those unsolvable circumstances (Yep, contrary to the junk science commonly taught in public schools, it is not a truly fixed constant - and neither is the speed of light, actually.)
The bottom line is, that while you can get GPS based PTS to work, you must allow several hundreds of feet extra margin for safety. You must also provide a fully functional coded audio frequency track circuit type signal system, as a basic foundation, in order to make it work at all, for failsafe train separation purposes. Moreover, the use the GPS data, is strictly limited to allowing the train to position itself when between wayside beacons. So it does not soley rely upon remembering the last wayside beacon - nor does it rely soley upon the GPS.
The big invisible elephant in the PTS room, and why I object; is that a specific vendor's technology has been mandated by political fiat, which has zero service record (not Amtrak's ITCS already in service, btw), The 2nd even bigger elephant is that conventional signal systems must still underlie the full use of PTS. However, this far cheaper conventional underlying system is in itself fully capable of all the safety sought after - without any so-called PTS at all.
So while it may be nice to have all the extra data in the cab that a GPS based system would afford, it is not essential to safety (the stated purpose of the mandate); so should not have been mandated. The free market could then decide if all the 'nice-to-haves' were worth the cost. What should have been mandated at a minimum, was the use of any system that failsafely and continuously supervised braking, while enforcing a not-optional stop prior to passing any stop signal.
Moreover, That mandate should have been applied based upon potential severity of ANY accident - in other words, maximum territorial speed - perhaps anything 30mph or over, unless permitted traffic in a whole territory was limited to only one train at a time.
But enough already - I'm still employed!
OPB
