tree68Just one more parameter to consider when trying to write a program that will handle potential situations.
I think I was successful in getting the R10 critical-systems group at ITU to understand the importance of tracking ad hoc situational awareness as well as just having the system monitor all the outputs from sensors and so forth. We had the critical advantage at that point of understanding the TMI post-mortems (and some of the Chernobyl machinations) and understood that a critical-response setup might in fact not at all mirror a control-room environment, particularly with respect to responses with complex interactions and intended consequences.
I'd like to think that we have a better or at least wiser set of systems programmers as the folks on NAJPTC who hard-coded a zero train length into the software at one point... considerably worse than not flexibly accommodating stopping short of formal control points. A problem is that the "Internet paradigm" for testing seems to be increasingly adopted, particularly by 'outsourced' bazaar-programming teams: the idea of going live bugs and all; your users will identify the bugs for you and notify you of what needs to be patched. It is my diplomatically-phrased opinion that this should never, ever be applied to critical systems, except as part of continuous optimization processes.
I had this discussion "again" with someone in Britain recently regarding lightweight rail vehicles and SPADs (not the airplanes but the near-unthinkable formal safety violations). It was and is my opinion that the brakes on these things need careful antilock operation, but also careful pre-application of KNOWN and UNKNOWN slip conditions in every modulated stop, applied and then monitored 'early enough' that time between stops can be minimized while absolutely ensuring no platform 'overrun'. Note that this implies some very sophisticated "autonomous" control in a system explicitly under 'driver' control at all times; one which should inspire no complacency that 'the system will handle it' without periodic 'stick time' as in some aircraft operations to maintain flying proficiency. Personally I think it's fun to design these things. Programmers who may be on a for-hire or consulting gig... perhaps not so much.
Overmod-- for example, I was once coming south on I-65 just past Fort Knox at night, and had the bottom fall out of a cloudburst on me in the middle lane at 65mph with no advance warning. Instantly there was zero effective vision and substantial undrained water on the pavement; the best I could do was to put on flashers, maintain some reasonable trajectory where others wouldn't hit me, and try to get to the right
Had essentially the same thing happen to me on I-44 at St. Louis once. At night. I was in the far left lane (of three). I didn't dare speed up, slow down, or change lanes even, as I couldn't tell if anyone else was around me, also flying blind. Even the lane markings were essentially invisible. It was only the occasional glimpse of the "Jersey barriers" to my left that kept me on course.
As I mentioned before, every trip is different. We regularly deal with leaves on the rails (just as bad as grease) and I once had one heck of a time with nothing more than the morning dew on the rails. Just one more parameter to consider when trying to write a program that will handle potential situations.
Larry Resident Microferroequinologist (at least at my house) Everyone goes home; Safety begins with you My Opinion. Standard Disclaimers Apply. No Expiration Date Come ride the rails with me! There's one thing about humility - the moment you think you've got it, you've lost it...
JohnMannYou have no idea how true that is! Automation does not understand and has no judgment.
Cheap 'automation' built to a price... with nothing but typical FSM algorithms and system complexity... yes, I agree with you. But if you look at autonomous-vehicle research since the understanding that the same principle that made the iPhone screen practical can be applied to massive sensor fusion, you will rapidly understand that haptic comprehension and effective safe judgment of actions have become easy to model and increasingly practical to implement. They just won't be simple-minded and deterministic.
For example, there was a program at Phantom Works in St. Louis using some of the logic derived from battlefield management systems like the extended version of Ida (which was 'born' as a detailing system for the Navy) which could track the progress of an air combat environment and determine when control inputs were dangerous (to the airframe, projected track, tactical situation, etc.) and take over the fly-by-wire system to produce alternative control outputs accomplishing 'pilot intent' by what might be highly nonstandard effectuation. You can probably deduce some of the necessary subsystems to accomplish this effectively, with combat-pilot input, in a secure and gracefully-degradable context. That is little different from what has become cost-effective to implement in 'production' autonomous vehicles.
What many people seem to forget is that autonomous vehicles will be smart and moral enough not to go into known bad weather conditions, or accident situations, or continue operating when being 'abused' by others in traffic, etc. They will do the equivalent of put on the flashers, go to the right, and if necessary navigate to safe parking or find alternate routes.
I have been in situations where all the prediction in the world might not help 'in time' -- for example, I was once coming south on I-65 just past Fort Knox at night, and had the bottom fall out of a cloudburst on me in the middle lane at 65mph with no advance warning. Instantly there was zero effective vision and substantial undrained water on the pavement; the best I could do was to put on flashers, maintain some reasonable trajectory where others wouldn't hit me, and try to get to the right. A proper autonomous vehicle would have weather-radar tracking in known poor-weather conditions, individual-wheel antilock braking, and sensors capable of resolving other traffic -- if necessary, by radio -- to avoid as much collision risk as possible; it would also 'know' the vehicle response characteristics to make the response essentially as good as 'humanly possible'. And note that nothing more than that could be expected of an automated system. We won't be able to prevent all accidents -- but we can avoid the usual ones, at least mitigate the unavoidable ones, and above all prioritize life in the responses. That is little different from what a human driver would do... if they had very fast reflexes, the ability to anticipate their situation or surroundings, be aware of and combine many information sources, revise schedule and assess JIT forward tracking and communication, etc.
Keep in mind that the system we designed for Conrail in 1987 (in the wake of the Chase wreck) included 'human-derived' power modulation and braking response to set up a given freight consist and bring it to a least-distance stop. In those days we had to hard-code a set of alternatives; nowadays we can accomplish much of what's needed with a simple set of running tests at and just out of the originating terminal... as well as having the hard-coded genetically-optimized rules and algorithms for anticipated train behavior.
BaltACD Believe corporate doubletalk at your own risk.
Believe corporate doubletalk at your own risk.
You have no idea how true that is! Automation does not understand and has no judgment.
SD70Dude Trip Op can't use the air brakes. For the longest time it couldn't even be programmed to obey throttle notch restrictions (it is supposed to now). CN doesn't use Leader, but we are supposed to use Trip Op whenever possible. But they aren't as serious about investigating people for not using it these days, not sure if the shine has worn off that idea or if they have simply laid off so many managers that there is no one left to monitor this sort of compliance and then conduct statements. Will Leader try to take air sooner and condition the brakes in winter, or does it think every day is warm and dry? Does it know where crossings are? Every big railroad claims that getting rid of conductors and eventually engineers will be the greatest thing since sliced bread. For now it's just talk to make themselves look good in front of the shareholders, as the systems have a long, long way to go before truly being able to run trains on their own.
Trip Op can't use the air brakes. For the longest time it couldn't even be programmed to obey throttle notch restrictions (it is supposed to now).
CN doesn't use Leader, but we are supposed to use Trip Op whenever possible. But they aren't as serious about investigating people for not using it these days, not sure if the shine has worn off that idea or if they have simply laid off so many managers that there is no one left to monitor this sort of compliance and then conduct statements.
Will Leader try to take air sooner and condition the brakes in winter, or does it think every day is warm and dry? Does it know where crossings are?
Every big railroad claims that getting rid of conductors and eventually engineers will be the greatest thing since sliced bread. For now it's just talk to make themselves look good in front of the shareholders, as the systems have a long, long way to go before truly being able to run trains on their own.
Our Trip Op system shows a place for fuel conservation (throttle notch) restrictions, but no way to engage it. It does have a condition speed, but that's something different.
CSX was supposed to be getting a new version of T-O called, "Trip Optimizer Zero to Zero." It's supposed to be capable of starting and stopping a train. I was hoping to find out on another forum that had sub forums for all the class ones how it was working out on CSX. Unfortunately it has been shut down.
Neither of our systems can automatically use air brakes. They prompt the use, and the release. Usually, both could avoid the use by just getting into dynamics, and staying in dynamics instead of waffling between dynos and power, early enough and hard enough that they wouldn't need to prompt for the use of air. Supposedly, the automatic control of air is close. It sounds like T-O may be there, but I haven't heard that my company is buying the new version. If my company isn't interested, it must be really bad if they're not interested.
When either system prompts for setting air, it's almost always just a minimum set. Rarely on my territory will it prompt to increase the set. We've always been instructed to not release until at least a 10 psi reduction had been made. It used to be a rule requirement until they changed it to allow EMS to release minimum sets. Now we have instructions to increase any EMS prompted reductions to at least 10 psi before releasing. I guess they were getting too many sticking brakes from releasing those minimum sets. (Theoretically, a minimum set-6 to 8 psi-is supposed to be enough so brakes won't stick on when released. In the real world 10 psi is about the minimum needed to avoid sticking brakes.)
I had techs from the railroad and LEADER ride with me once. That we don't always go right up to the stop signal is a concept they don't understand. That we may stop way short of the end of authority to stay off crossings, etc. doesn't "compute" with them. At least in their offices/cubicles/laboratories.
Jeff
SD70DudeTrip Op can't use the air brakes. For the longest time it couldn't even be programmed to obey throttle notch restrictions (it is supposed to now). CN doesn't use Leader, but we are supposed to use Trip Op whenever possible. But they aren't as serious about investigating people for not using it these days, not sure if the shine has worn off that idea or if they have simply laid off so many managers that there is no one left to monitor this sort of compliance and then conduct statements. Will Leader try to take air sooner and condition the brakes in winter, or does it think every day is warm and dry? Does it know where crossings are? Every big railroad claims that getting rid of conductors and eventually engineers will be the greatest thing since sliced bread. For now it's just talk to make themselves look good in front of the shareholders, as the systems have a long, long way to go before truly being able to run trains on their own.
The ultimate aim is to get rid of everybody except the CEO so he can push the botton on the money machine to the stockholders. And then the stockholders will cry for the CEO's head.
Never too old to have a happy childhood!
Greetings from Alberta
-an Articulate Malcontent
Most of our fleet of modern AC engines are equipped with energy management systems, the auto throttle. If available, we have to us it 96% of the trip to be compliant. They were having some resistance, so upped it to 96% from 93%. If not used, the reason why must be submitted.
The other day I had it disengage itself twice. It became available again after about 30 miles. The second time I never saw it become available again. It's integrated into PTC and for some reason, some engines fitted with the screen had them place it so one has to turn their head to the left about 95 degrees. That's some way to place a safety device.
We've had a lot of the auto throttle break trains into two or more pieces. LEADER is better than Trip Optimizer at train handling. Both sometimes do stupid things, but T-O does it more often. So much, that I consider T-O nothing more than a speed control. It does whatever it wants to reach the speed the system has projected.
A while back, one of my coworkers went to an investigation for train handling. He had a break in two, but that wasn't the reason. Anytime there is an incident, they have been reviewing the tape (black box download) for the entire trip. We have specific guidelines for certain areas and if you don't follow them, you can be disciplined. This engineer was running the train the way he had been taught by the old heads. Unfortunately, the specific guidelines often are written by those who haven't run trains except on a simulater, where they come up with the guidelines. EMS is exempt, they can run the train any way it wants.
Preparing for the investigation, our local chairman asked the railroad for data about break in twos and other train handling issues. How human engineers compare to the EMS auto throttle. The railroad said they don't collect data on EMS incidents, it's treated as a mechanical failure.
At least that's what I was told when LEADER tore up my train a few years ago.
Euclid BaltACD Euclid Murphy Siding How do they handle the braking situation like what just happened in Kansas City? It depends on what the situation in Kansas City was. But bringing a train down a grade would not pose any fundamental problem for autonomous operation. Most idiots can get a train up a grade - it takes an ENGINEER to get today's size trains down the grades. There is much more involved in getting todays sized trains down grades than getting them up the same grades. Nobdy is disputing that. What does it have to do with the ability of autonomous train driving systems to take trains down a grade?
BaltACD Euclid Murphy Siding How do they handle the braking situation like what just happened in Kansas City? It depends on what the situation in Kansas City was. But bringing a train down a grade would not pose any fundamental problem for autonomous operation. Most idiots can get a train up a grade - it takes an ENGINEER to get today's size trains down the grades. There is much more involved in getting todays sized trains down grades than getting them up the same grades.
Euclid Murphy Siding How do they handle the braking situation like what just happened in Kansas City? It depends on what the situation in Kansas City was. But bringing a train down a grade would not pose any fundamental problem for autonomous operation.
Murphy Siding How do they handle the braking situation like what just happened in Kansas City?
It depends on what the situation in Kansas City was. But bringing a train down a grade would not pose any fundamental problem for autonomous operation.
Most idiots can get a train up a grade - it takes an ENGINEER to get today's size trains down the grades. There is much more involved in getting todays sized trains down grades than getting them up the same grades.
Nobdy is disputing that. What does it have to do with the ability of autonomous train driving systems to take trains down a grade?
Autonomy is dumb. Dumb is not SAFE.
My hubby used to say this about some of his cargo. He said that it was safer to haul a tanker half full on ice than a load of swinging beef in a crosswind. Why the tanker was lower to the ground than the beef was and didn't tend to literally rock and roll in the wind.
EuclidWhat does it have to do with the ability of autonomous train driving systems to take trains down a grade?
"Seat of the pants" is an important part of running a train.
IF all of the parameters that are feeding the software are correct, and assuming that the software is capable of knowing the territory in question foot by foot, no problem. But it's a well known bit of wisdom that no two trips, even with exactly the same consist, are exactly the same.
If any of those parameters are out of whack, and the program doesn't have the answer, chaos will ensue.
Most idiots can get a train up a grade - it takes an ENGINEER to get today's size trains down the grades SAFELY. There is much more involved in getting todays sized trains down grades than getting them up the same grades.
EuclidBut bringing a train down a grade would not pose any fundamental problem for autonomous operation.
I know several hoggers that had the autothrottle/leader tear their train into several pieces.
It's been fun. But it isn't much fun anymore. Signing off for now.
The opinions expressed here represent my own and not those of my employer, any other railroad, company, or person.t fun any
Murphy SidingHow do they handle the braking situation like what just happened in Kansas City?
I wasn't sure which of the autonomous rail car threads to add this to, so I just picked the newest one.How would they handle switching the cars at each end and all the sidings and yards in between? Most of our lumber comes from Canada with a lot of miles, a border, a railroad change and several switching yards along the way. There's been talk about fleeting(?) of the cars once they're out on the main.Would the cars be hooked together into a train with 100 inividual engines to control?How do they handle the braking situation like what just happened in Kansas City?
Thanks to Chris / CopCarSS for my avatar.
Euclid daveklepper Full automation is possible with todat'sc technology only in completely closed systems, where no random occurabce can impinge on the operation. And, outdoors, weather itseklf can be a souce of a serious random occurance in many parts of North America. Grade crossings and shipper erors in loading are also seriousw causes. Today,, there are automatic ralways: People movers at airports and specific rapid -transit lines, the latter almost always with an operator on board for security. No grade crossings, and single track only when one or two trains (midway passing siding) are used. Intra plant freight could be another application and dedicated single origine and single destination, with no weather problems, in abarren isolated areas, another. Dave, I agree with your point about the need to avoid random occurrences that impinge on the operation. This is major obstacle to the development of autonomous cars and trucks on roadways because hazardous impingement can come from any direction at any time. If it does, the driver’s attention is likely to recognize it as an approaching threat, and react defensively in a way that avoids the danger. So roadway autonomous driving requires extremely thorough sensing and program ability to deal with the threat of things like tires coming off of trucks, or out of control vehicles originating from other lanes or even from other roadways. This issue is practically non-existent with autonomous trains. Much of the potential impingement will not damage a train, and even if it could, trains are not likely to be able to stop or slow down enough to prevent the collision. Yet this issue of grade crossing hazards is widely promoted as the reason why railroads absolutely cannot adopt autonomous operation. I think that objecting is overstated. There are cases where an engineer can spot a dangerous obstruction on a grade crossing and react in time to mitigate or even avoid the collision. For that capability, autonomous trains will be well equipped with forward sensing that will detect any obstacle in the path of the train. As I understand Rio Tinto operation, their trains follow a program that correlates certain actions with the train’s location on the line. But they do have automatic sensing specifically to verify that grade crossings are clear. All 6 or so of them on the automated line. And most in remote locations as I understand. Other than this need for intense sensing ability and perfection, autonomous trucks and cars are practically fully developed and ready for use. They may be awaiting full regulatory approval, and require further testing to achieve that. However, railroad autonomous trains with their fixed guideway principle, do not pose the challenge of providing the degree of intense sensing that is required for vehicles being constantly steered and traveling on roadways where the impingement potential is very critical. Designing the application for railroads will be the main challenge because the operation is so complex. But fundamentally, I think Autonomous trains are nearer to implementation than either autonomous trucks or cars. My perception is that the railroad industry wants to implement this autonomous technology. If they want it, the providers will certainly provide it. There could not be a more attractive market for the developers than the railroads with their deep pockets and massive market size arising from their standardization. The supply industry and its creators are always trying to find new products to sell to the railroads because of the attractiveness of that market. But because the railroad industry is so besieged with new ideas, they are very selective and skeptical. ECP brakes are a great idea, but the railroads do want it because they do not see it as offering a cost effective solution. I think the situation is the complete opposite with autonomous trains. For one thing, railroad management tends to feels that autonomous trains are necessary to compete with autonomous trucking. For another thing, the driverless trains eliminate labor cost. Bingo! We have a winner! eliminate labor costs. Reading Rio Tinto's goals is to automate everyting involved in the mining process, except for maintenance.
daveklepper Full automation is possible with todat'sc technology only in completely closed systems, where no random occurabce can impinge on the operation. And, outdoors, weather itseklf can be a souce of a serious random occurance in many parts of North America. Grade crossings and shipper erors in loading are also seriousw causes. Today,, there are automatic ralways: People movers at airports and specific rapid -transit lines, the latter almost always with an operator on board for security. No grade crossings, and single track only when one or two trains (midway passing siding) are used. Intra plant freight could be another application and dedicated single origine and single destination, with no weather problems, in abarren isolated areas, another.
Full automation is possible with todat'sc technology only in completely closed systems, where no random occurabce can impinge on the operation. And, outdoors, weather itseklf can be a souce of a serious random occurance in many parts of North America. Grade crossings and shipper erors in loading are also seriousw causes.
Today,, there are automatic ralways: People movers at airports and specific rapid -transit lines, the latter almost always with an operator on board for security. No grade crossings, and single track only when one or two trains (midway passing siding) are used. Intra plant freight could be another application and dedicated single origine and single destination, with no weather problems, in abarren isolated areas, another.
A320 (A319/A321) Auto Brakes
When I read about the automation of trains, I remember my trip on a PATCO train to & from Lindenwald NJ. This was back about 60 years ago. The trains are "semi-automatic" in that they were programed to run with an operator who pushes a button to close the doors and another to start the trip. The train then accelerates to track speed and runs to the next station where it stops and the doors open. Repeat to the next station. All day long. On my trip, the operator pushed the buttons at each station on the outbound trip but on the inbound trip, he manually controlled the train using the throttle and braking into the next station. When we got closer to Philly, I asked him about it. He told me it was to maintain his proficency, as the automatic braking was just based on a track point and that with the programed braking rate, if there was wet rail, the train would slide and overshoot the platform. This was of course in the early days of "automation" but I saw a similar case where I was told that SWA did not want its pilots using some of the automatic landing systems because they wanted thier pillots to NOT loose their proficencies. This has been an issue for some pilots who fly the long routes and may only land a plane five times a month.
We're a long way still from full automation, if that goal even makes sense from a cost/benefit standpoint. Look at how difficult and costly it was to get PTC up and running. Full automation will likely require significant advances in artificial intelligence to where the robots on board are able to think as well as a human being can. Apart from recent formidable advances in robotics and neuroscience, we've only scratched the surface.. we're probably many decades away from understanding how our minds work let alone how to emulate our intelligence in technology.
NYAB, who makes the LEADER system is one of the groups involved with Rio Tinto's automation. We run LEADER integrated with PTC. LEADER once got me a PTC enforcement.
We had stopped on the downhill portion of one of our steepest grades. We met a train and got a signal to go. I started the train. When prompted for auto available, I gave it control. We were in a 45 mph zone with a 40 mph zone about a mile plus ahead of us. Additionally, I had the condition speed set to 40 mph due to a cold weather restriction.
When I gave it control we were doing about 25 mph and in dynamics. It promptly got out of dynamics and went into power. We were on a sharp curve and I was looking back as we went around it. It had control for about half a mile when I turned back and looked at the screens.
We were doing 41 mph (According to the loco's speedometer. PTC showed 40 mph and LEADER uses the PTC speedometer. Trip Optimizer uses the loco's speedometer.) still in power and accelerating. I was reaching for the automatic air brake handle when PTC gave me a 2 second warning for the approaching 40 mph zone. It then took our air. We stopped, everything reset and we went on our way. I reported the enforcement as required, giving all the details, upon tie-up.
I never heard anything more about it.
If Rio Tinto's system is so great, why isn't everyone else scrambling to license their technology and implement it right away?
Euclid He does not go into any of the issues involved with introducing autonomous freight trains. He just sounds like he is opposed to them. Why? He seems to say the concept is only used on rail transit and not on freight trains. Has he heard of Rio Tinto?
I don't agree. I don't think he's against it or that it won't come. I think he's focused on the flippant attitude of those who really don't know how things, in this case railroads, work. That we can just flip a switch and, voila, robot trains, and that's what will solve the supply chain problems.
I see the same thing in the hreads on autonomus freight vehichles. It's not a question of whether it can be done. It's a question of practicality. Techies always think the answer to a problem is more technology. It's all they know. They don't always know or understand how things work. I once had 2 railroad techs and 1 NYAB tech ride with me when they were rolling out the auto throttle version of LEADER. It was clear all they knew about train operations was it went from point A to point B. They didn't understand that on that journey, if the train had to stop that sometimes, most of the time, the train shoudn't go all the way up to a red signal. That there are other considerations in train operation besides just going from A to B.
As to Rio Tinto. Maybe he does know. Maybe he knows better on how well it's working, instead of relying on a press release on how it's the greatest thing ever. I would like to know how well it's working. I'm sure it works well enough, but there is a failure rate and I'd like to know what it is. I'd like to hear about some of the failures.
This is reminiscent of the schemes in 'steam days' used on the LMS in Britain and at least one railroad in France to track fuel consumption carefully and encourage its economical use.
As I recall, the English system was punitive in nature (i.e. consequences if you failed to meet norms)whereas the French provided both 'egoboo' and some kind of financial reward for performance. In my opinion the latter is preferable if substantial fuel savings with reasonable morale is the object.
I was nodding along until I got to the part about student engineers learning to power-brake. By the time I was sure I understood the euphemistic circumlocutions my hands were starting to clench to strangle somebody.
Our community is FREE to join. To participate you must either login or register for an account.