Positive Train Control and High Speed Rail

[Jersey_Mike]

Bah, LZB's a better system.

LZB requires additional wires to be installed between the rails which would be a no go here.

Works on the Canarse line, and it'd work a lot better if the NYCTA would get everything equipped on the line.

That's debatable. From what I've heard the CBTC has been nothing but problems and the only saving grace is that they left the existing signaling system in place so there's immediate failover to the old, perfectly functional system. Even if they do get it working it is unclear if the whole enterprise will fail the cost benefit test. At least for the real railroads nobody seems to be proposing that track circuits be replaced for train detection.

In any case, the 10 or so trains a day the average main line in the US sees, a cab signal system isn't going to even have any effect on running times. There's CSS equipped lines with 10 times the traffic out there in the world that work just fine.

The main lines that matter are the ones that see 50 trains a day, which is an ever increasing share of them. Besides the issue isn't CSS, its some newfangled wireless system. Basic 4 aspect CSS would have most likely prevented the Chatsworth crash. There is no need for PTC, no need for computers and wireless datagrams. The mandate should have been for the defacto national standard 4 aspect pulse code CSS. It's bullet proof and has a consistent track record dating back to the 1930's. Every one of the big 4 Class 1's use it somewhere on their system and all Amtrak locomotives are equipped.

BTW do people realize that once they start going wireless they are going to have to secure the system and I mean really secure the system in such a way it doesn't get blown out of the water at Defcon. Wireless means that any joker with a laptop and $30 bux worth of equipment can start sending false speed target to trains. The technology isn't hard, but to make a secure system railroads are going to have to deal with, bum bum bum, key management for the message authentication codes. They have to key every PTC unit in ever cab then re-key it probably monthly. Of course you know they'll never do this so at least if my Amtrak train is going slow, after 2015 I can just command it to go faster.

Port Authority Transit Corp.'s thoughts on the Lindenwold, N.J. High Speed Line were that a human operator is often more conservative than the Automatic Train Operation because the human might practice "defensive driving" to avoid being fired if the train doesn't stop in as short a distance as the computer thinks it should. Going down hill, a Lindenwold train coasts and increses its speed until it exceeds the signal system's speed limit, when the brakes automatically slow it down and then release. A human must run slightly slower to keep the train under the speed limit at all times, or possibly have to explain why not, at the risk of his job.
Washington Metro is another example. It's Automatic Train Operation just applies the brakes at a predetermined point (based on speed, location, and other factors) without worrying about overshooting the stopping point and hitting another train.
Are you suggesting that Postive Train Control computers will be programmed to run trains slower than past practice?

On both PATCO and the DC Metro you are correct that the ATO runs the fastest possible time as constrained by the signaling system. However the comparison I am making is with a signalling system that is augmented by human judgement instead of hard and fast speed rules. Locomotive engineers use their own judgement on where to begin to brake based on route knowledge, signal indication, etc. So a freight train that passes an approach medium signal will begin to slow down sooner than an Amtrak train that can maintain 80mph right up to the interlocking then brake to 30mph in the last 2000 feet. All of the computerized systems have large safety margins built in so when you transition to them the system slows down. An experienced engineer is going to be able to operate his train with less margin for error than a computerized system will. Yes this sometimes results in mishaps, but I feel that is a worthwhile tradeoff given the amazingly few number of accidents that occur each year.

If you don't believe me look at at highway speed limits and advisory speed signs. Traffic engineers have all sorts of very sophisticated equations to determine what the target speed on any particular stretch of highway should be and we all know how much total BS they are. Flow of traffic on 55mph freeways is 70 and unless you're driving a 1960's Ford Galaxy you can take the cloverleaf ramp marked 25mph at 50mph. Imagine if the government required you to have a little box in your car that forced your car to actually obey all of those speeds. The railroads have already had the engineers show up and impose their conservative speed limits on everything, now they risk taking away the leeway the engineers had in following the track and signal speeds.

[NellieBly]

Jersey Mike:

We probably ought to just take this discussion private. But I do want to say (to you and the others interested enough to post here) that I've been working with communications-based control systems now since 1987. I rode BN's test installation on Minnesota's Iron Range two decades ago, and the train handling was every bit as good as a human engineer could do, and this was a 22,000 trailing ton train of iron ore.

There is no reason whatever to expect the train handling algorithms in PTC will be more conservative than a human engineer. The assertion by AAR that train-specific braking algorithms aren't possible because train consists aren't accurate enough is tendentious nonsense. If a train consist is incomplete (especially if it fails to include hazmat cars) the railroad is in violation of FRA regulations. AEI readers exist at hundreds of locations; these can validate consists.

PTC represents a sea change in train control philosophy. Four-aspect cab signals are an obsolete, 80-year-old technology. As for their being "bulletproof", tell that to the nine dead people in the June WMATA accident. And wayside-centric technology like that, in addition to being unreliable and obsolete, is far more expensive than PTC. The cost of ETMS is so staggering that even the cash-rich government railroads in Europe are complaining. PTC will cost a tenth as much.

[Jersey_Mike]

PTC represents a sea change in train control philosophy. Four-aspect cab signals are an obsolete, 80-year-old technology. As for their being "bulletproof", tell that to the nine dead people in the June WMATA accident. And wayside-centric technology like that, in addition to being unreliable and obsolete, is far more expensive than PTC. The cost of ETMS is so staggering that even the cash-rich government railroads in Europe are complaining. PTC will cost a tenth as much.

Stop talking about PTC like its is a specific technology, PTC means positive stop and other speed restriction enforcement which includes everything from trip arms to ETMS. The DC Metro has PTC (not 4 aspect cab signals) and it didn't stop the crash. In fact the crash was caused by a failure of train detection, not the signaling logic. You seem to be implying that the "future" is train control will be full on CBTC, well what do you thing ETCS (phase III) is? You're going from one relay hut every 2 miles to a transceiver every 500 feet with all the quirkiness of wireless data transmission (can you hear me now?) Why do you think CBTC costs so much and takes so long to install? The only advantage is full moving block, which is only ever needed in very high capacity metro systems. I also love the assumption that GPS is some sort of reliable utility like the electric grid. If there's a huge solar storm or major military conflict are we going to lose the ability to signal our trains?

Incidentally, can you explain how PTC is radically different from ETCS? ETCS makes use of internal navigation based on track mounted beacons with movement authorities transmitted wirelessly to the train. That's basically what you said "PTC" will do except it uses GPS instead of beacons.

GPS based systems as you described might work well as a replacement for current DTC/TWC/DCS territory with no signals and dictated track authority, but those have 10 mile long blocks and don't need trains to stay in constant contact with anything. However is this really a good alternative for a typical automatic block system with track circuits? What's going to ensure that hand operated switches have been left closed or that a freight car hasn't rolled out onto the main line? If you keep the track circuits then where's the huge cost savings? Remember the runaway boxcar on the NEC? The continuous CSS got the MBTA commuter train stopped while "Virtual blocks" would not have detected that.

The fact that you work with CTBC only confirms your bias. Of course you would promote a technology that will guarantee you full employment for the rest of your life. Regarding BN's test installation, unit iron ore trains exhibit uniform handling characteristics (just like passenger trains). On the regular railroad trains can have cars of different weights all mixed up in any order. While I am sure they could tweak the algorithms to match the handling profile of a single type of unit train, that doesn't mean it will work in the rail network at large. I mean if this was really a cost saver don't you think the Class 1's would have done it already? They aren't stupid. They've been replacing code lines with microwave links and using remote control locomotives in yards so its not like they are being ultra conservative here.

Just because something is old, doesn't mean it's bad. The IT industry has been frothing at the bit to be able to go in and replace a few cheap relays with complex computer hardware that requires constant and expensive support. You want to see the result? Go to Philly and try riding the downtown Trolley Tunnel. Amtrak actually did a good job with the ACSES PTC system that's almost fully deployed on the NEC. If there's a failure they have the old CSS as a fallback and it also relies upon the CSS for the dynamic updates of track condition. No fancypants wireless datagrams that can get blocked by someone turning on their microwave or spoofed by a hacker.

This isn't to say that something using GPS and internal navigation won't be used to replace DCS et al on unsignaled lines or very low traffic lines that need to be re-signaled anyway (Southern Tier, R&N Lehigh Line, etc), but on main lines that need track circuit detection the way to go is either a transponder based system backed up by wayside or cab signals in case a transponder fails.

So once again I am interested what exactly will qualify as a PTC system because something that looking like ETCS will probably suffer the same fate as ETCS.

[AFTower]

Amen,Jersey_Mike! Amen!

[neroden]

Class 4+ track maintenance requirements is the biggest hurdle to high speed rail.

In fact, there's a fair amount of "supermainline" track which is already maintained at Class 5 so that fast freight can go at 79 mph instead of 60 mph. When the now-mandatory PTC systems are installed, the class 5 track will mean an *immediate* increase in passenger train speeds from 79 mph to 90 mph.

Don't expect railroads to install anything that requires them to re-signal or re-interlock their existing routes.

They will -- to a certain extent. There are an awful lot of track-warrant-controlled, unsignalled main line segments which they will have to install signalling on thanks to the mandate. This will be the dominant factor in the choice of signalling system, and as such they'll choose a cab signalling system. As long as they're putting cab signals in all the locomotives, they will do their best to save money by getting rid of waysides everywhere, which means partial resignalling. (Interestingly I suspect this means there will be a set of "shortline/branchline locomotives" which are banned from mainlines due to lacking cab signals, and which must be towed when transported from one branch to another.)

However, I assume the underlying train detection and block systems will remain in place on multi-tracked lines and at major interlockings. GPS-based signalling simply doesn't work at all for double-track lines, sidings, or anything where the tracks are too close together, because it's not precise enough. But the procedure for communicating the signal aspects to the trains is going to change; cab signalling is the way of the future, and waysides are just expensive by comparison.

I would expect fairly conventional cab signalling to be used on double-track lines, with speed authority communicated through the tracks, simply because, as of now, it's too hard for any other system to tell whether it's talking to the train on track 1 or the train on track 2, and getting that wrong is a recipe for crashes. I suppose a GPS-based system might be used in rural single-track areas, with the signal communicated through the track replacing the GPS signal in multitrack areas (signalling "You are on track 1 in block N. You are on track 1 in block N." over and over).

ANYway, this unsignalled mainline track also has a 59 mph passenger speed limit currently, but has a freight speed limit of 49 mph, so will often be class 4 (since class 3 has a freight speed limit of 40 mph). Class 4 has a passenger speed limit of 80 mph, so signalling on these tracks will bring the speed limit up from 59 mph to 80 mph.

Finally, the signalling has been the main expense preventing railroads from raising track speeds in some spots. After the signalling is done, the economics may cause them to increase class 5 track to class 6 or class 4 track to class 5, in places where it was previously not worth it because the signallng-based speed limits prevented the maintenance increase from actually increasing freight train speed limits.

So the PTC mandate is no panacea but it will probably bring passenger speed limits up on a lot of tracks -- 79 to 90 on class 5, 59 to 80 on class 4. That is nothing to sneeze at.

[tarheelman]

Class 4+ track maintenance requirements is the biggest hurdle to high speed rail.

In fact, there's a fair amount of "supermainline" track which is already maintained at Class 5 so that fast freight can go at 79 mph instead of 60 mph. When the now-mandatory PTC systems are installed, the class 5 track will mean an *immediate* increase in passenger train speeds from 79 mph to 90 mph.

Don't expect railroads to install anything that requires them to re-signal or re-interlock their existing routes.

They will -- to a certain extent. There are an awful lot of track-warrant-controlled, unsignalled main line segments which they will have to install signalling on thanks to the mandate. This will be the dominant factor in the choice of signalling system, and as such they'll choose a cab signalling system. As long as they're putting cab signals in all the locomotives, they will do their best to save money by getting rid of waysides everywhere, which means partial resignalling. (Interestingly I suspect this means there will be a set of "shortline/branchline locomotives" which are banned from mainlines due to lacking cab signals, and which must be towed when transported from one branch to another.)

However, I assume the underlying train detection and block systems will remain in place on multi-tracked lines and at major interlockings. GPS-based signalling simply doesn't work at all for double-track lines, sidings, or anything where the tracks are too close together, because it's not precise enough. But the procedure for communicating the signal aspects to the trains is going to change; cab signalling is the way of the future, and waysides are just expensive by comparison.

I would expect fairly conventional cab signalling to be used on double-track lines, with speed authority communicated through the tracks, simply because, as of now, it's too hard for any other system to tell whether it's talking to the train on track 1 or the train on track 2, and getting that wrong is a recipe for crashes. I suppose a GPS-based system might be used in rural single-track areas, with the signal communicated through the track replacing the GPS signal in multitrack areas (signalling "You are on track 1 in block N. You are on track 1 in block N." over and over).

ANYway, this unsignalled mainline track also has a 59 mph passenger speed limit currently, but has a freight speed limit of 49 mph, so will often be class 4 (since class 3 has a freight speed limit of 40 mph). Class 4 has a passenger speed limit of 80 mph, so signalling on these tracks will bring the speed limit up from 59 mph to 80 mph.

Finally, the signalling has been the main expense preventing railroads from raising track speeds in some spots. After the signalling is done, the economics may cause them to increase class 5 track to class 6 or class 4 track to class 5, in places where it was previously not worth it because the signallng-based speed limits prevented the maintenance increase from actually increasing freight train speed limits.

So the PTC mandate is no panacea but it will probably bring passenger speed limits up on a lot of tracks -- 79 to 90 on class 5, 59 to 80 on class 4. That is nothing to sneeze at.

Thanks for the informative post. I hope you're right about the immediate speed limit increases once the signalling is addressed.

[Jersey_Mike]

In fact, there's a fair amount of "supermainline" track which is already maintained at Class 5 so that fast freight can go at 79 mph instead of 60 mph. When the now-mandatory PTC systems are installed, the class 5 track will mean an *immediate* increase in passenger train speeds from 79 mph to 90 mph.

I'm not holding my breath. There's a lot of 70mph track out there. It's class 4 so why not 80? I'm been told that every additional mph presents more maintenance issues so PTC is not going to instantly bring about high speed rail.

They will -- to a certain extent. There are an awful lot of track-warrant-controlled, unsignalled main line segments which they will have to install signalling on thanks to the mandate. This will be the dominant factor in the choice of signalling system, and as such they'll choose a cab signalling system. As long as they're putting cab signals in all the locomotives, they will do their best to save money by getting rid of waysides everywhere, which means partial resignalling.

I'm all for Rule 562 operation with full home signals and distants, but Conrail seemed to be the only Class 1 freight railroad that really embraced it and we all saw the way SEPTA's flawed implementation slowed down operations. Depending on how flaky this CBTC system is look for the wayside signals to stick around, even the plain ABS ones. Also keep in mind that the PTC's job is to keep trains safe by enforcing various speeds via penalty brake applications (which may or may not piss away the air and cause a runaway or derail the train due to slack action), but not ensure proper train handling to avoid a penalty dump. Engineers sometimes need to act on a signal before they pass it even to avoid a penalty brake application so this is where waysides would come in handy. The cost is with the track circuits and many thousands of miles have just gotten new wayside signals in the last 10 years. Newly resignaled lines might go 562, but we'll have to see.

(Interestingly I suspect this means there will be a set of "shortline/branchline locomotives" which are banned from mainlines due to lacking cab signals, and which must be towed when transported from one branch to another.)

What's more interesting if this will mean the end of rare millage passenger excursions or if tourist lines will have to deal with this crap.

I would expect fairly conventional cab signalling to be used on double-track lines, with speed authority communicated through the tracks, simply because, as of now, it's too hard for any other system to tell whether it's talking to the train on track 1 or the train on track 2, and getting that wrong is a recipe for crashes.

If you don't use some sort of track mounted beacon or coded track circuit you are just asking for mis-labeling accidents to occur. I think the railroads would just love to not have to install any wayside devices, just try to get it to work as an installed cab module, but I just don't see that being feasible.

ANYway, this unsignalled mainline track also has a 59 mph passenger speed limit currently, but has a freight speed limit of 49 mph, so will often be class 4 (since class 3 has a freight speed limit of 40 mph). Class 4 has a passenger speed limit of 80 mph, so signalling on these tracks will bring the speed limit up from 59 mph to 80 mph.

Read the FRA definition of "signaling system" and unless they changed things many non-track circuited manual block systems count as one. The NORAC rulebook states the maximum speed in DCS territory to be 70mph, not 59. Also be aware that while some main lines may be built to Class 4 standards, they are still run at 60 or even 40mph. If the line is hilly and doesn't have many high speed segments they would have to flag any Class 4 defect and put out a speed restriction so its just easier to drop the speed to 60 or 40. C&O main line is 60mph. B&O is 40-50 east of Cumberland. C&O track was immaculate, but it was still 60mph.

If I had to rank the impediments to high speed rail in order it would be mountains, maintenance and then signals a distant 3rd.

[neroden]

Read the FRA definition of "signaling system" and unless they changed things many non-track circuited manual block systems count as one.

Bleh, that's right.... Practically, however, Amtrak has repeatedly stated that they are limited to 59mph in "dark territory".

The NORAC rulebook states the maximum speed in DCS territory to be 70mph, not 59. Also be aware that while some main lines may be built to Class 4 standards, they are still run at 60 or even 40mph. If the line is hilly and doesn't have many high speed segments they would have to flag any Class 4 defect and put out a speed restriction so its just easier to drop the speed to 60 or 40. C&O main line is 60mph. B&O is 40-50 east of Cumberland. C&O track was immaculate, but it was still 60mph.

If I had to rank the impediments to high speed rail in order it would be mountains, maintenance and then signals a distant 3rd.

I've been looking through this, and you seem to be largely right. There are a number of lines which were Class 5 in the recent past and seem to have been allowed to drop to class 4. Even on lines which are Class 5 the speed limit often seems to be below 79 (the Chicago-Aurora line, for instance). I swear some of those major, flat, straight intermodal lines are going to end up with 90 mph speed limits pretty quickly though.

I'm having trouble spotting what Amtrak trains still run in dark territory. Do any of them, except for the Vermont trains (which run on track with very low speed limits anyway)? I know the old Sunset Limited east did. Probably the Cardinal, but again I think it runs on tracks with very low speed limits.

[Jersey_Mike]

I'm having trouble spotting what Amtrak trains still run in dark territory. Do any of them, except for the Vermont trains (which run on track with very low speed limits anyway)? I know the old Sunset Limited east did. Probably the Cardinal, but again I think it runs on tracks with very low speed limits.

The Cardinal runs DCS between Crawfordsville and CP-CLERMONT on the old Cawrfordsville, Secondary in Indiana. The three Rivers used to run DCS on the old Youngstown Line from CP-ROCHESTER to CP-75. Amtrak also has Rule 251 on the Harrisburg Line and on some NEC tracks so you need Form D's on those for reverse running as well. Anyway, true "Dark Territory" is like timetable and train order operation with superior and inferior trains.

BTW another thought I had was that in dark territory the new PTC stuff will automate movement authorities, but there will still be no train detection so if something happens on the back end server to "lose" the train one is going to have a problem. Nobody should trust commodity "software" for vital safety applications. Stick with relays, ROMs and ASIC's. Once you have SignalSoft 2.0 sending vital safety to trains in the field you know its going to be running on Windows and you know its going to get connected to the internet and that spells Trouble with a capitol T and that rhymes with P and that stands for PTC I'm a computer security researcher and the best way for "industry" to try to keep SCADA systems secure is to force someone to actually drive out into the field and muck with it. If you make it accessible to the network then state or terrorist actors can sit back and completely pwn you from the safety of their home country. Centralized Computer based interlocking is all the rage in Europe. Hopefully they will be providing the world with a nice little object lesson on the ways that can go horribly wrong before someone tries to adopt it over here.

[justalurker66]

Anyone who hates railroads is going to love a system that saddles them with an expensive burden and ties the system in knots while Joe the Trucker and the Bolt Bus fly down interstate 80 at 80mph door to door.

Besides that you assume failure of PTC on railroads (are you one who hates railroads and wants them to fail?) what you suggest on the highway is illegal. Are you suggesting that safe railroad operations be thrown out the window so railroads can compete with unsafe road operations?

An experienced engineer is going to be able to operate his train with less margin for error than a computerized system will.

That is the same problem with road jockies. Joe the Trucker gets a little experience and over confidence and decides that it is safe for him to violate the law by doing 80 on 80, while eating a sandwich and text messaging. Then he rounds a curve he's rounded hundreds of time before, glances down to adjust his radio and looks up just in time to see stopped traffic backed up from a construction zone. Joe kills a bunch of people.

While the Chatsworth passenger engineer was the most wrong for going through red signals without paying attention the Chatsworth freight engineer was also text messaging while operating his train. Both displayed the same kind of experience based overconfidence with their job.

Positive Train Control can be many things. It can be as simple as an alert system to tell the engineer that things are not as they should be. From something as simple as the crossing gates he expects to be working a mile ahead have not triggered properly to letting him know that some train ran a stop signal and is fouling his track. There is no reason why trains would be run any slower than they are today - UNLESS THOSE TRAINS ARE BEING OPERATED ILLEGALLY.

I hope you're not suggesting trains should be operating illegally. Are you?

[justalurker66]

Once you have SignalSoft 2.0 sending vital safety to trains in the field you know its going to be running on Windows and you know its going to get connected to the internet and that spells Trouble with a capitol T and that rhymes with P and that stands for PTC I'm a computer security researcher and the best way for "industry" to try to keep SCADA systems secure is to force someone to actually drive out into the field and muck with it. If you make it accessible to the network then state or terrorist actors can sit back and completely pwn you from the safety of their home country.

Closed networks would be helpful. While the last connection between the network and the train would need to be protected there is no need to put the system on the internet. Use some of that ROW to run a private network connected to your CTC center.

BTW: If you're a computer security researcher why are you not pushing Linux or an embedded operating system that is (allegedly) more stable and less vunerable? Or are you one of those MS certified hackers that thinks MS is the only way and one should pay for all of the MS security patches plus pay MS experts to break their own work to create the next round of security patches and continue the cycle until Bill Gates' foundation has enough money to buy the world?

[Jersey_Mike]

I hope you're not suggesting trains should be operating illegally. Are you?

No, I'm suggesting that the computers will cause trains to operate with a larger cushion below the posted speed limits in order to ensure that they never accidentally go above them. As it is right now all track and turnout speeds already have a margin of error built in so if an engineer takes a 30mph turnout at 40mph they won't derail (I've seen the AAR turnout ratings, a #15 (30mph) is officially good for 42mph and a #10 (15mph) is good for 28-30mph) and same with speeds for curves and whatnot. PTC runs the risk of putting a margin of safety on top of another margin of safety!!! Engineers have the skill to run their trains in a way that they can get right up to the line and occasionally they will inadvertently go over, but that has already been taken into account by the transportation-type engineers who set the speeds in the first place.

Besides that you assume failure of PTC on railroads (are you one who hates railroads and wants them to fail?) what you suggest on the highway is illegal. Are you suggesting that safe railroad operations be thrown out the window so railroads can compete with unsafe road operations?

I wasn't aware that railroad operations were unsafe as they are now. Railroads are already at a competitive disadvantage because they actually follow the rules. How is it going to help to make the rules 20% stricter? As soon as you have a device that's designed to keep trains "safe" you have a magnet that will attract liability paranoia. After every lawyer at US&S and the railroads dip their fingers into the PTC we'll be lucky if they'll be able to move at restricted speed.

Closed networks would be helpful. While the last connection between the network and the train would need to be protected there is no need to put the system on the internet. Use some of that ROW to run a private network connected to your CTC center.

Closed networks are never closed, especially in the commercial IT world where there is every incentive to plug computers in to the internet because its 100 times easier to get software updates and other tools onto the box that way. Also people are notoriously weak when it comes to being able to resist constant gMail access.

BTW: If you're a computer security researcher why are you not pushing Linux or an embedded operating system that is (allegedly) more stable and less vunerable? Or are you one of those MS certified hackers that thinks MS is the only way and one should pay for all of the MS security patches plus pay MS experts to break their own work to create the next round of security patches and continue the cycle until Bill Gates' foundation has enough money to buy the world?

Pushing it for whom? You suggesting I should be nagging people on this board to use Linux? For PTC I'm taking things a step beyond saying that the system shouldn't even use a non-embedded OS and it most certainly shouldn't use long range wireless communications. I know people are going to tell me to relax and suggest that the signaling folk will make something that's secure, but I've seen far too many failures to trust any industry except financial services and the DoD.

[justalurker66]

There are a few years to work it all out ...

[Jersey_Mike]

There are a few years to work it all out ...

I'm hoping/assuming that the result will be something along the lines of TPWS in the UK which was able to be fitted to the large numbers of mechanical semaphores still in use (and even installed new in places). Now I don't mean that specific technology, but something that can be quickly bolted right to the existing wayside or cab signal system and when it's not working right doesn't send trains crawling around at restricted speed. We've seen one version of the the future and it is SEPTA. They took 5 years to install CTBC on its transit tunnel which still doesn't work right and is only being used for political reasons. Hell, NJT knew it had a real stinker with the SES system on the PVL and they weren't able to dump it until the line was re-signaled under Rule 562 two years ago.

Remarkably, Amtrak's has a winner with the ACSES overlay. As a rider I don't even know it's doing anything as speeds didn't get slowed down since they installed it and there aren't issues with trains getting stopped due to the new stuff failing. They discovered the whole idea with the stop release combination codes didn't work well so they dumped it. NJT's adopting it so we'll probably see it on Metro North, MBTA and the LIRR to fulfill the PTC requirements.

[justalurker66]

One thing for sure is it won't be one thing. It will be what it needs to be for each railroad.