Reading the text it seems to me this is due to a very widespread issue in Spain and other southern European countries: people don’t care about the warnings because there are too many false positives.
That’s why people here, when they hear the fire alarm they just wait for it to go off: the probability that it is actually a fire is minuscule… but the alarm triggers a lot.
Not sure how to actually solve this… but I’m also not making excuses for what happened. High speed trains is not the same as a fire alarm in some office building, the people in charge should be held responsible somehow.
It just means that a false positive rate is as dangerous as a false negative rate. Just issuing a warning doesn't come for free when you are designing a security system.
Something similar is happening in my region after the flood that happened two years ago: The alarm was raised late; Now, even mild rain is triggering an alert.
Guess whats the consequence of that? Less and less people pay attention to the alerts.
What you mention makes me frustrated, as it also signals the infrastructure is so underdeveloped even moderate rain grinds daily life to a halt: kids don't go to school if rain is forecast. We need to elevate our professionalism in Spain and not accept African-level infrastructure in the 21st century. It's really sad within one continent we have countries producing AI and photon computing, and others can't even figure out how to put gutters and rain drains.
> That’s why people here, when they hear the fire alarm they just wait for it to go off
This literally happened to me last night, I was having dinner at my apartment, and suddenly I start hearing a smoke alarm, but not quite inside the apartment, I just kept having dinner thinking "bah, someone probably burned a toast".
Then it wont stop, I walk to door and hear people talking, still don't bother enough, wait a few minutes, then go out and my neighbor said "there are three fire trucks down there, I think it's a real one" hahhaha
The hallway was kind of hazy and smelled like burnt food, for sure... so I just grabbed my laptop and went out hahaa.
Because they don't get held responsible enough for bypissing procedures. Some system are too complex for fully automated operarions, and the human need to play along.
Safety culture is blameless, but only if everyone plays along.
The level of human negligence here is astounding. The system did exactly what it was supposed to (lock the train down if the track can't be confirmed clear), and then a human controller ordered the crew to override the system, and ordered them to proceed at an unsafe speed towards a similarly disabled train?
Whose negligence, though? "The system" isn't just the LZB equipment - the train dispatcher is just another cog inside the same system, and should be expected to occasionally fail just like every mechanical part.
This isn't a game of Hot Potato where the last person touching it goes to jail for gross negligence. You have to look at it through a Swiss Cheese model: every individual safety layer can and will have holes, but a disaster can only occur when all of those holes align.
Why did the LZB system fail? Why did the train dispatcher use the wrong backup procedure? Why did a backup procedure with zero safety checks even exist? Why did the train operator accept this backup procedure? Why did the train allow high-speed driving under those unsafe conditions?
It does indeed seem like the train dispatcher made a big mistake, but that was far from the only thing that went wrong, and it would be myopic to use him as the scapegoat for a far larger systemic issue.
Dispatcher, conductor at that time could not have know that dispatcher ordered manual bypass ignorig the rest of the guardrails of a traffic resume procedure
I would argue that we should look at systems-level explanations. The human factor is so glaring we can get caught up in it. There are what, 50?, high speed trains in movement in Spain at the same time. Even GPS + Starlink + 1 monitor (is that 50k€?) could have prevented this. But it happened, and the question is what is going on with train security in Spain in general for people to (be able to) make such decisions.
Trains are still not able to assess properly where they are, even using satellites and inside train sensors. Of course there is research on that, but it's still an unsolved problem. (Current alternative is to put markers in the track, but that's also expensive at scale.)
Yeah and that's fucking ridiculous and a damning indictment of the glaring incompetence of train companies.
If I can have reasonably accurate gps readings of my phone while I'm on the train, there is zero reason for train companies to not have systems that are at least as accurate.
> a damning indictment of the glaring incompetence of train companies.
No, what I describe has nothing to do with the companies, that's the research state. It's just an indictment of you making bold statement about a subject, you don't now enough about.
> I can have reasonably accurate gps readings of my phone
Would you dare to cross on a rope over a deep valley, blind using only GPS readings of your phone? Because that's kind what you are suggesting, expect you also move a few hundred other people.
The way I think of it, divide the tracks into segments. Before you put a train on a segment, you must ask a question - is there a train on it. There are only two answers possible, No, and something else. If it is something else, stop. Don't let the train on that segment.
There are optimizations possible like make segments smaller etc but the original algorithm is trivial. If we can't figure things out, we make segments longer and we slow down the trains until we can. I am positive for a sufficiently large track segment and a sufficiently slow train, we can always answer the question with a no.
> Would you dare to cross on a rope over a deep valley, blind using only GPS readings of your phone?
Hang on, this is nothing at all like that - trains run on fixed tracks. The only bit of data we are interested in for collision avoidance is the linear offset along the track. Even the shittiest GPS fix of all-time would give a good-enough intersection with a fixed path to avoid crashing into a second stationary train in the open country-side (do you see any satellite line-of-sight blockers in the crash footage?)
> The only bit of data we are interested in for collision avoidance is the linear offset along the track.
So that is naive, but assuming that would be all that you need, with GPS you would also need the positions of all other trains. But why would you transmit all this information to every train, instead of just computing the collisions in a central node? (That's what is done now.) A better fit for what you suggest would be radar, and this does exist. [1] But that's where your original assumption is naive. You regularly pass other trains within <2m at high speeds, so linear offset is not all you need.
> But why would you transmit all this information to every train, instead of just computing the collisions in a central node?
Because central nodes can fail (as they did here, both human and machine), and providing this redundant data allows both the computer and the human-in-the-loop on the fast-moving chunk of steel to independently validate the decisions coming from the central node.
> You regularly pass other trains within <2m at high speeds, so linear offset is not all you need.
In this collision, both trains were on the same track. So while we obviously care about adjacency issues in the general case, this specific failure mode can be detected purely via a linear offset.
You can and do have redundancy in the central node already. Also that's not what failed here.
> In this collision, both trains were on the same track.
Yeah, but how do you distinguish that? So when I understand you correctly, what you want is: Find another moving chunk of metal in the radius of 2km, which moves toward you, but not in a direct or ballistic trajectory, but in a arbitrary complicated trajectory, but which can lookup in a database. Also you are operating in the real world, so there are potentially houses, trees and mountains in between. I guess that would be possible, but it would be really expensive. Your best bet would be to adapt a missile detection and targeting system from the military. A system which does exist and is somewhat similar is RCAS (https://www.dlr.de/en/kn/research-transfer/projects/next-gen...).
I feel like you are intentionally trying to turn this into a spherical cow situation by insisting on solutions that work in an information vacuum. All of this sort of problem in the real world is solved by sensor fusion from multiple data sources.
Trains are captive to a particular track - they can only change tracks at specific pre-determined junctions. With a pretty minimal sensor feed from each junction we can statically determine which track segment a given train is on. With GPS we can shrink that segment length down to a few kilometres. By sharing data persistently every single train can validate that central command-and-control is not about to run it into another train from behind...
There cases which are genuinely challenging engineering problems (i.e. derailments that impinge on adjacent tracks), but this trivial case of bumper cars is not one of them.
How do you distinguish that the <2km away detected moving train runs on your track on not on the one directly besides you?
> Trains are captive to a particular track - they can only change tracks at specific pre-determined junctions.
Yeah, but you do not know which one it runs on. Your entire point is that you do not trust the information the infrastructure knows. If you do allow to use information from the infrastructure, then why not use the sensors that already exist along the whole train network? It's also very unlikely that there is something electrically connecting the two rails of a track, that you would like to run over. So why wouldn't you trust that information, but some vague GPS signal?
> we can statically determine
But the current state of the art is, that you deterministically and provable know where a train is. Why would you use a more expensive, less safe system? The current focus to reduce errors in computing train conflicts is to prevent random bit flips due to radiation at the same time in three different computers running different software.
> I feel like you are intentionally trying to turn this into a spherical cow situation
The situation where you proposal makes sense, but the current system does not, is a spherical cow situation, that's my point.
> Trains are still not able to assess properly where they are, even using satellites and inside train sensors. Of course there is research on that, but it's still an unsolved problem
Throwing up our hands about how hard a problem this is seems like a ridiculous position to stake. If we can design self-driving cars, that operate safely in far less controlled environments, we can figure out the linear offset of a large vehicle traveling along a physical rail running along a fixed pre-determined path.
I said nothing about "throwing up out hands", this is a current research topic. But you can hardly blame the workers for not using a system, the researchers haven't invented yet.
> If we can design self-driving cars
We can not design self-driving cars to the safety standard that is required for railways. They are way less safe than the current half-manual system for railways. What you suggest is using less automatic safety systems than what we currently do.
Has there ever been a case where two self driving cars collide with each other? I hate to be part of the peanut gallery, but as a totally naive person when it comes to railway safety research it really does seem amazing train collisions can’t be easily prevented with gps, lasers, sonar or various other trivial ways. Even the crash in question seems like it might have been easily prevented by having a second human in the loop. Maybe I and others just suspect the difficulty isn’t so much technological as it is financial.
As far as I am aware, there has been no non-supervised train in a general train network. (They exist in closed systems.) However whether the train is self-driving is really not the issue here. In current railway safety systems (even from the 80s) a crash is about the same probability as random bitflips. However reality is too messy, so that the availability is not satisfactory enough to not use a human fallback, and then you are back to the assisted human error rate.
Current operation of a train is basically self-driving and supervising driver. So if you suggest "just use self-driving", you make the system less safe, not more, because it means just getting rid of the human supervisor. This is not currently done, because it would be too insecure. It might be done in the future, but rather because less safety is going to be accepted in favor for getting rid of the costly human.
A self-driving train will happily cause a crash, if the target is "drive with high-speed in that segment currently occupied by another train". The issue is that that becomes a target, not that there is a human in the loop.
Would you mind sharing some more of the why it is still an open problem? (I know I know little about this subject.) I understand satellite positioning isn't definitive and did not suggest it should be the /primary/ risk mitigating factor. But even as a tertiary or lower risk mitigating factor, this would have been a worthwhile investment. (Sorry for the bleak introduction of economics.)
And some problems are just _open_ to outside looks. My assistant tells me that the room for error in trains on one dimension ('left - right') is mere 10s of centimeters. I understand that problem. But what track you are on, isn't that obvious to the human in the cabin? (Again, simplifying, not flaming. Production ready safety is _hard_.) So could a vision based solution help count tracks from left to right? (Assistant: "US PTC, on about 54,000 route-miles, uses GPS for the lead locomotive plus crew-entered track information and switch position".) So you use the dimension that matters most in crashes (forwards / backwards) from satellite, augmented with vision or human based left-right positioning. Some trains already have sensors to sense things like leaves on the track to adjust speeding and braking... So the sensor based approach is already there.
Perhaps I am read as suggesting that the problem is easy, and that is not my intention. But for cars this was solved decades ago for routing and years ago for self-driving. I use that daily. Flying is organised with a very clear 6 sigma approach. Crashes are nearly always situations where events / risks stack. Trains (in Spain)...?
Tl;dr - Safety is a function of (human; process; technology) and there is clearly something lacking for trains, when compared to other types of transport.
So that's a lot of questions, that I all can't answer in full.
> Would you mind sharing some more of the why it is still an open problem?
Because for train sensors, there is a small drift which accumulates too much, hence it is synchronized by magnetic position markers in the tracks.
> I understand satellite positioning isn't definitive
Yes, and the point is it might be some solution if we would have the safety systems from 1860. But we do not and hence it needs to be more accurate than the current system, and it is not and in addition also more expensive (and subject to remote influence and reliance on potentially adversary nation states).
> Sorry for the bleak introduction of economics
No, economics is basically the primary constraint for safety. We could pay million people to make sure a single person does not stumble if we want total safety, but that's completely pointless, and also these million people would starve, because no one is producing food anymore.
> But even as a tertiary or lower risk mitigating factor
But for that it's too expensive and still unreliable. There is RCAS[1], which I already linked elsewhere in the thread. That's basically cutting edge and too expensive to introduce everywhere. Also previous safety systems are already not rolled out enough, which is causing these accidents, so rolling out yet another expensive system is not going to improve much.
> But what track you are on, isn't that obvious to the human in the cabin?
Yes, but if it is already known, why use GPS to calculate the same information with more error?
> So could a vision based solution help count tracks from left to right?
If the train driver or dispatcher makes use of vision to assess the train position, the error will be higher than the current sensor/electric-resistance based system.
According to Wikipedia, ASFA being introduced in 1975 is rolled out in the entire Spain network. ETCS which is the European standardization seems to be rolled out on all high-speed lines or currently being rolled out.
> So you use the dimension that matters most in crashes (forwards / backwards) from satellite, augmented with vision or human based left-right positioning.
Duh, that just sounds like trains would be operating blindly in space currently and need to assess their approximate position.
> Perhaps I am read as suggesting that the problem is easy, and that is not my intention.
No, the issue is that you are suggesting far more complicated solutions to already solved problems, with simpler systems in operation for decades. The current system is that nobody in the train even has a say on where the train goes, so the problem you are trying to prevent doesn't even exist. This IS why it's way more safer than a car. There are movement authorities, that define from where you are allowed (and have to) drive with your train. The problem is not, that the train driver isn't able to know where he is and hence drives wrong or too far (that would be prevented by the on-board safety system), but that the movement authority itself is faulty.
On an up-to-date signaling box (withing the past 6 decades), even the train dispatcher is no say in safety questions. You could put a monkey in charge of it. There would be chaos, but they would not cause a crash. The current job of a dispatcher in normal operation is to assign resources to prevent deadlocks and to implement vaguely defined priorities. (Basically the equivalent of the public thinking a programmer is a code monkey, while the actual point is clarifying stakeholder interests.)
The actual issue here is, that while these safety system have a very small false positive rate, they do have a too large false negative rate for there to be an acceptable availability. Hence in the latter case, the human needs to take over and then you are back to a much higher false positive rate. The issue would be alleviated by relaxing the constraints on cost per availability. Either by allowing for less availability or by reintroducing more humans, so that the system situation can be assessed by humans directly.
These accidents are caused as follows: the automatic system fails, because the internal state of the system does not match the actual state of the world. Hence the human needs to take over. But the human also needs to rely on some system to know the world state. Which systems can he trust, given that the whole situation is that the systems report that he can't trust them? So he needs to make some decision and in most cases, that decision is still correct. But there is still an error rate to that, and then something like what we are discussing happens.
> But for cars this was solved decades ago for routing and years ago for self-driving.
This is absolutely not solved for the cars, and to the sense that it is, it was solved and rolled out from the 1870s onward for railways using mechanical computing systems, the switch to using electricity happened in the 1910s.
> Flying is organised with a very clear 6 sigma approach.
You are correct, that aircraft safety systems need to be designed to a higher reliability than railways, because on a rail you can always stop in case of an error, while in the air such a choice doesn't lead to a safe state. But if you compare these safety systems, basically aircraft is operated like railways used to be operated in the 1920s. (My peers actually had a task to compare them and find ideas from one that can be applied to the other. While there is a lot of automation to be taken from rail that can be applied to the aerial industry, there is not really much that can be applied the other way.) It is also operated with the amount of personal than 1920s railways. That's because there is a lot of money in aircraft, while railways are mass-transit and need to be cheap. If you would operate today's railway with the amount of human oversight than in the past, it would be a lot safer than aircraft.
> there is clearly something lacking for trains, when compared to other types of transport.
What are you talking about? Railways are so much safer than cars. https://news.ycombinator.com/item?id=50288518 Yes, there are less accidents in planes than in trains, but this is due to what I already explained.
The current control system dates from the early 90s. They started rolling a modern system with more redundancies and automatic blocking in 2022 but it's still not fully deployed.
PZB is from 1934. Arguably always prioritizing deploying the most modern system on the highest profile lines, is what prevents us from properly rolling out the basic safety systems everywhere.
I was only talking about the general problem domain, not about this specific location. The current safety system is also not was has caused this, likely. Or rather it has caused this by it's availability, not by it's safety.
Sorry, but ordering trains to drive manually after the system locks itself down, is the proper procedure. Of course you are still supposed to only send trains to free sections, but the error is in that step, not that they ordered a train to drive on it's own.
Also the fallback procedure is also aided by automatic systems and a sizable number of accidents in the past were caused by that information being faulty.
It is the proper procedure, and it does come (in a lot of countries, AFAIK) with a lot of guardrails:
There will be a so-called "train order" allowing passage of a specific section ignoring the safety system warnings, which has strict validity limits both in time and in the area where it applies, usually denoted by specific hectometric mileposts
When cab signaling is involved (LZB is one of these) and it malfunctions, the entire system normally changes to "driving on sight only" - which indeed limits the train to 40, 30, 20, sometimes - 15 km/h (see https://fr.wikipedia.org/wiki/Marche_%C3%A0_vue) or the line gets switched to lineside signaling - if it is available.
I don't know how you could violate all of those rules - both from the driver's side and from the traffic control side - but it sounds absolutely horrible. Like an override of at least 2 (if not 3) very explicit safety margins.
I remember a long time ago I was stuck in a train in a field a considerable time because of a similar failure and conductor just wouldn't move unless green
I was expecting the same here, conductor is the first to die so just refuse until the idiots fix their system
A conductor always just gets a movement order. They are not able to assess whether the track is clear, that's the reason why the signalling system exists in the first place. How they get that movement order doesn't necessarily tell what the security steps were to back it.
Of course an attentive conductor can notice and mention there something being weird ("Doesn't the other train normally arrive first?"), but that's a rare lucky occasion, nothing to be relied on. And they just need to swallow the answer ("No, you go first today...") anyway.
> and conductor just wouldn't move unless green
The conductor won't ever move unless they get a movement authority, which most of the time is a green signal.
Why would they trust "a random blinking green light", but not a written and stamped order? Yes, they could do that (and sometimes it will actually prevent an accident), but it will be more being a snowflake, then being a decision for safety. A conductor is simply not responsible for whether it's safe to drive on a given track. Their equivalent is not a car driver.
The "random blinking green light" is generally more reliable
Being a snowflake here has 2 benefits:
1. Getting company managing tracks to fix the lights (they need to deal with angry customers and pay off damages for delays and more), so there's a direct path from unreliable lights to economic damage.
2. Avoid crashing and killing yourself and passengers on the offchance the person on the radio is a moron
> The "random blinking green light" is generally more reliable
Yes, but not because of the light itself, but due to the safety steps done before. Which should also have happened before the movement order in any other form.
As a train driver, you do not have the information required to distinguish these situations. If drivers simple refuse all movement orders, besides "green lights", running trains will become a rare occasion and the drivers will simply get fired for not doing their job. _The train driver is not the one who decides where a train goes._ For this safety-relevant decision a driver is (roughly speaking) nothing more than an innocent bystander.
You can think of a train driver more of the person who makes sure the automatic engine works correctly and also has an eye, that nothing weird happens. Most of your train driver education is knowing how the train works and how to get a broken train back working fast. In addition you also learn the safety protocols how you get the information in which state the engine should operate in. This is what train signals are, it's not really the equivalent of a green traffic light for car traffic.
I would actually sit in front now. Sometimes you would be able to still get vital information and being actually able to act. But the safe decision would be no to enter the train.
It's incredible the minister didn't resign after the January crash and I doubt he will now either. I understand that perhaps the problem started in the control centre but surely at the end of the day the minister is responsible - they don't get paid a fortune just to pass the blame on to the workers.
I fail to see how the two accidents, which seem to have very "down to earth" explanations (maintenance failure by workers, protocol override by operators) are being assumed to be a political responsibility so quickly by some and asking for the minister resignation. Now, two in a row seem to point to systemic issues. Is this responsibility of Adif or Renfe? maybe those state companies have to be thoroughly revised. is it due to lack of funding? maybe taxes should be raised, or money diverted from somewhere else (though seeing the spanish budget, it's hard to tell where from, both teachers and doctors are currently protesting for lack of funding). Maybe it is, in the end, a political directive that should have responsibilities declared in the executive. But seeing so many interested parties jumping to that conclusion literally hours after the crash means it's not done out of any reasoning but political animosity
> are being assumed to be a political responsibility so quickly by some and asking for the minister resignation
It sucks but that's how it is over here. It's "my" party the one in power when a catastrophe like this happens? It's not the politicians, it's someone else. Is it the "other" party? Surely the politicians are ultimately responsible.
I can't imagine what every passenger who survived and the families of those who passed are going through. Such a catastrophic incident with multiple human failures is so tragic and so needless. From the controller to the driver, every single one of them should have said hey we shouldn't just ignore a signal and should investigate why it's on first. For this to be such a quick decision suggests it's not a one off. I would be extremely interested to see how many near misses or ignored signals there has been across Spain. The crash earlier this year was due to inexperienced workers and it seems as if significant training improvements and protocols are needed.
To have had this many incidents compared to other HSR networks around the world gives all the signs of a classic case of normalisation of deviance.
Sadly we'll probably watch a lot of individuals being blamed/scapegoated for systemic issues, but hopefully even Spanish politics can organise a proper investigation after this many deaths in a single year.
If as a driver you get told by the control centre to get the train moving again at maximum line speed, surely you should be able to trust that it is safe to do so?
I don't see how this would fall on the drivers at all.
My thoughts as well. It appears that such signal failures happen so often that the controllers regularly let trains run through red under verbal authorization. I don't know the exact protocols, but I would expect a driver to ask for confirmation, as happened here, and then trust the control centre. However, there is clearly a huge systemic problem, with added complexity for the controllers, and there should at least be double checks in the control centre.
The person who controls whether the light shows red or green and gives verbal authorization is the same. It's both a tool, to transmit the same information.
Trust but verify. You wouldn't drive your car into a lake just because your satnav tells you to, would you?
As we also know from the aviation industry, controllers are human and therefore will make mistakes. A controller shouldn't issue a wrong order, but a driver also shouldn't accept an order they aren't certain about, and there should be technical systems backing both sides of that up.
It doesn't "fall" on the driver, but it doesn't "fall" solely on the controller either. It's a systemic issue, and many pieces have to fail for a disaster to happen.
The aviation industry operates like the railway industry ~1920 (somewhat by necessity).
> You wouldn't drive your car into a lake just because your satnav tells you to, would you?
You are assuming a train driver is equivalent to a car driver. He's not, the analog would be more a repairmen that you always take with you on journeys.
> From the controller to the driver, every single one of them should have said hey we shouldn't just ignore a signal
A train driver has zero insight in what is going on and where other trains are. They do need to trust the orders they get, if you want to question every order you will never start to drive.
No, he would get fired, because that's not his job. It would be like the machine engineer on a race track refusing to fix the ignition, because somewhere on the track he doesn't see, the asphalt might be broken.
The point of this discussion is that people blame the train driver or say "use self-driving", when that is not going to solve the problem, because that's not where the error is. The error is not there, *because* for rail making errors operating the train is not going to cause an accident (most of the time), which is not the case for cars.
The driver always has the final say on whether the train is safe to move. They are not just a button and lever pusher. They have to analyse whether the decision is correct. Hence why I stated there were multiple points of failure. The controller and the driver both went with full speed despite it being completely against regulation on a red signal.
I agree. But the decision whether all switches are locked in the correct position and that the track is clear, is not one that he is supposed to make and he also doesn't have the information for it.
> The controller and the driver both went with full speed despite it being completely against regulation on a red signal.
Going full-speed in the case of a red signal is not necessarily against regulation. Also the information of "red signal" or "you can go" is both a signal from the same source. There is no information for the driver to make a different decision over the dispatcher.
I can't comment about whether the current accident was caused by going against regulation, and more important by whom and against which. We need to wait for the proper safety report, not for a BBC article. If you take the latter at face value, it would mean there are clowns operating, which I am sure, is not actually the case. The real error is always way more subtle, otherwise it would not have happened.
> The train driver's job is to say: "it's red, I'm no moving, not my problem, you make it green if you want me to go"
A written movement authority is a signal as well in railway terms. It's more like "I don't trust your second traffic light, that you placed next to the normal one, because the normal one is broken".
The even more insane thing to me is that the controller knew that there was another train in that block, knowing it was no signal failure. The driving train should’ve never been commanded to continue on full speed by the controller.
Trains shouldn’t even be manually driven in 2026. We have self driving cars now which is a way harder problem to solve. Relying on a system of signals and driver attentiveness is archaic.
> Trains shouldn’t even be manually driven in 2026
Potentially, but the point is that signallers should physically be unable to allow trains to proceed, and drivers should physically not be able to override them. British signalling from the 1800s has physical interlocks; this isn't a matter of manual versus automatic train operation. It is clearly systemic.
We do not have self-driving cars in all environments that clear the necessary security regulation! Also the socially accepted number of deaths due to trains is way below that for cars. The current accident is just a Tuesday for car traffic. (Spain recorded 1785 road deaths in 2024[1], that's 4.9 per day) Detecting objects (and future objects!) on a train track is still a hard problem.
The current system is basically that the train is driven both automatically and by a professional attentive driver at the same time. Changing it to just the automatic system is going to reduce safety. Also the real world is way to messy to reach acceptable availability rates with only automatic systems.
Yeah, that's a solved problem. Did you even read my whole comment? I answered why "just use the system from self-driving cars" wouldn't work and in fact would be less safe than the current system.
Driving a train and driving a car are two completely different challenges.
If signaling worked properly then it wouldn't matter whether the train was driven autonomously or not, the "problem" becomes trivial. Establishing and continuously verifying that signaling is working properly is the hard part.
Without signaling you're limited to crawling speed just like humans are when they don't have any signalling .
Reading the text it seems to me this is due to a very widespread issue in Spain and other southern European countries: people don’t care about the warnings because there are too many false positives.
That’s why people here, when they hear the fire alarm they just wait for it to go off: the probability that it is actually a fire is minuscule… but the alarm triggers a lot.
Not sure how to actually solve this… but I’m also not making excuses for what happened. High speed trains is not the same as a fire alarm in some office building, the people in charge should be held responsible somehow.
It just means that a false positive rate is as dangerous as a false negative rate. Just issuing a warning doesn't come for free when you are designing a security system.
This is very well illustrated in the fable Boy who cried false positive.
that sounds like a character from Harry Potter and the Methods of Rationality.
Something similar is happening in my region after the flood that happened two years ago: The alarm was raised late; Now, even mild rain is triggering an alert.
Guess whats the consequence of that? Less and less people pay attention to the alerts.
What you mention makes me frustrated, as it also signals the infrastructure is so underdeveloped even moderate rain grinds daily life to a halt: kids don't go to school if rain is forecast. We need to elevate our professionalism in Spain and not accept African-level infrastructure in the 21st century. It's really sad within one continent we have countries producing AI and photon computing, and others can't even figure out how to put gutters and rain drains.
> That’s why people here, when they hear the fire alarm they just wait for it to go off
This literally happened to me last night, I was having dinner at my apartment, and suddenly I start hearing a smoke alarm, but not quite inside the apartment, I just kept having dinner thinking "bah, someone probably burned a toast".
Then it wont stop, I walk to door and hear people talking, still don't bother enough, wait a few minutes, then go out and my neighbor said "there are three fire trucks down there, I think it's a real one" hahhaha
The hallway was kind of hazy and smelled like burnt food, for sure... so I just grabbed my laptop and went out hahaa.
Because they don't get held responsible enough for bypissing procedures. Some system are too complex for fully automated operarions, and the human need to play along.
Safety culture is blameless, but only if everyone plays along.
[dead]
The level of human negligence here is astounding. The system did exactly what it was supposed to (lock the train down if the track can't be confirmed clear), and then a human controller ordered the crew to override the system, and ordered them to proceed at an unsafe speed towards a similarly disabled train?
Whose negligence, though? "The system" isn't just the LZB equipment - the train dispatcher is just another cog inside the same system, and should be expected to occasionally fail just like every mechanical part.
This isn't a game of Hot Potato where the last person touching it goes to jail for gross negligence. You have to look at it through a Swiss Cheese model: every individual safety layer can and will have holes, but a disaster can only occur when all of those holes align.
Why did the LZB system fail? Why did the train dispatcher use the wrong backup procedure? Why did a backup procedure with zero safety checks even exist? Why did the train operator accept this backup procedure? Why did the train allow high-speed driving under those unsafe conditions?
It does indeed seem like the train dispatcher made a big mistake, but that was far from the only thing that went wrong, and it would be myopic to use him as the scapegoat for a far larger systemic issue.
Dispatcher, conductor at that time could not have know that dispatcher ordered manual bypass ignorig the rest of the guardrails of a traffic resume procedure
I would argue that we should look at systems-level explanations. The human factor is so glaring we can get caught up in it. There are what, 50?, high speed trains in movement in Spain at the same time. Even GPS + Starlink + 1 monitor (is that 50k€?) could have prevented this. But it happened, and the question is what is going on with train security in Spain in general for people to (be able to) make such decisions.
Trains are still not able to assess properly where they are, even using satellites and inside train sensors. Of course there is research on that, but it's still an unsolved problem. (Current alternative is to put markers in the track, but that's also expensive at scale.)
Yeah and that's fucking ridiculous and a damning indictment of the glaring incompetence of train companies.
If I can have reasonably accurate gps readings of my phone while I'm on the train, there is zero reason for train companies to not have systems that are at least as accurate.
> a damning indictment of the glaring incompetence of train companies.
No, what I describe has nothing to do with the companies, that's the research state. It's just an indictment of you making bold statement about a subject, you don't now enough about.
> I can have reasonably accurate gps readings of my phone
Would you dare to cross on a rope over a deep valley, blind using only GPS readings of your phone? Because that's kind what you are suggesting, expect you also move a few hundred other people.
who said anything about only using gps? I'm saying in addition to their current system!
The exact comment I answered to. https://news.ycombinator.com/item?id=50287160
The way I think of it, divide the tracks into segments. Before you put a train on a segment, you must ask a question - is there a train on it. There are only two answers possible, No, and something else. If it is something else, stop. Don't let the train on that segment.
There are optimizations possible like make segments smaller etc but the original algorithm is trivial. If we can't figure things out, we make segments longer and we slow down the trains until we can. I am positive for a sufficiently large track segment and a sufficiently slow train, we can always answer the question with a no.
Congratulation you are reinventing 1890 state of the art.
> Would you dare to cross on a rope over a deep valley, blind using only GPS readings of your phone?
Hang on, this is nothing at all like that - trains run on fixed tracks. The only bit of data we are interested in for collision avoidance is the linear offset along the track. Even the shittiest GPS fix of all-time would give a good-enough intersection with a fixed path to avoid crashing into a second stationary train in the open country-side (do you see any satellite line-of-sight blockers in the crash footage?)
> The only bit of data we are interested in for collision avoidance is the linear offset along the track.
So that is naive, but assuming that would be all that you need, with GPS you would also need the positions of all other trains. But why would you transmit all this information to every train, instead of just computing the collisions in a central node? (That's what is done now.) A better fit for what you suggest would be radar, and this does exist. [1] But that's where your original assumption is naive. You regularly pass other trains within <2m at high speeds, so linear offset is not all you need.
> But why would you transmit all this information to every train, instead of just computing the collisions in a central node?
Because central nodes can fail (as they did here, both human and machine), and providing this redundant data allows both the computer and the human-in-the-loop on the fast-moving chunk of steel to independently validate the decisions coming from the central node.
> You regularly pass other trains within <2m at high speeds, so linear offset is not all you need.
In this collision, both trains were on the same track. So while we obviously care about adjacency issues in the general case, this specific failure mode can be detected purely via a linear offset.
You can and do have redundancy in the central node already. Also that's not what failed here.
> In this collision, both trains were on the same track.
Yeah, but how do you distinguish that? So when I understand you correctly, what you want is: Find another moving chunk of metal in the radius of 2km, which moves toward you, but not in a direct or ballistic trajectory, but in a arbitrary complicated trajectory, but which can lookup in a database. Also you are operating in the real world, so there are potentially houses, trees and mountains in between. I guess that would be possible, but it would be really expensive. Your best bet would be to adapt a missile detection and targeting system from the military. A system which does exist and is somewhat similar is RCAS (https://www.dlr.de/en/kn/research-transfer/projects/next-gen...).
> Yeah, but how do you distinguish that?
By enumerating the tracks?
I feel like you are intentionally trying to turn this into a spherical cow situation by insisting on solutions that work in an information vacuum. All of this sort of problem in the real world is solved by sensor fusion from multiple data sources.
Trains are captive to a particular track - they can only change tracks at specific pre-determined junctions. With a pretty minimal sensor feed from each junction we can statically determine which track segment a given train is on. With GPS we can shrink that segment length down to a few kilometres. By sharing data persistently every single train can validate that central command-and-control is not about to run it into another train from behind...
There cases which are genuinely challenging engineering problems (i.e. derailments that impinge on adjacent tracks), but this trivial case of bumper cars is not one of them.
>> Yeah, but how do you distinguish that?
> By enumerating the tracks?
How do you distinguish that the <2km away detected moving train runs on your track on not on the one directly besides you?
> Trains are captive to a particular track - they can only change tracks at specific pre-determined junctions.
Yeah, but you do not know which one it runs on. Your entire point is that you do not trust the information the infrastructure knows. If you do allow to use information from the infrastructure, then why not use the sensors that already exist along the whole train network? It's also very unlikely that there is something electrically connecting the two rails of a track, that you would like to run over. So why wouldn't you trust that information, but some vague GPS signal?
> we can statically determine
But the current state of the art is, that you deterministically and provable know where a train is. Why would you use a more expensive, less safe system? The current focus to reduce errors in computing train conflicts is to prevent random bit flips due to radiation at the same time in three different computers running different software.
> I feel like you are intentionally trying to turn this into a spherical cow situation
The situation where you proposal makes sense, but the current system does not, is a spherical cow situation, that's my point.
Would that work in tunnels and forests too?
> Trains are still not able to assess properly where they are, even using satellites and inside train sensors. Of course there is research on that, but it's still an unsolved problem
Throwing up our hands about how hard a problem this is seems like a ridiculous position to stake. If we can design self-driving cars, that operate safely in far less controlled environments, we can figure out the linear offset of a large vehicle traveling along a physical rail running along a fixed pre-determined path.
I said nothing about "throwing up out hands", this is a current research topic. But you can hardly blame the workers for not using a system, the researchers haven't invented yet.
> If we can design self-driving cars
We can not design self-driving cars to the safety standard that is required for railways. They are way less safe than the current half-manual system for railways. What you suggest is using less automatic safety systems than what we currently do.
Has there ever been a case where two self driving cars collide with each other? I hate to be part of the peanut gallery, but as a totally naive person when it comes to railway safety research it really does seem amazing train collisions can’t be easily prevented with gps, lasers, sonar or various other trivial ways. Even the crash in question seems like it might have been easily prevented by having a second human in the loop. Maybe I and others just suspect the difficulty isn’t so much technological as it is financial.
As far as I am aware, there has been no non-supervised train in a general train network. (They exist in closed systems.) However whether the train is self-driving is really not the issue here. In current railway safety systems (even from the 80s) a crash is about the same probability as random bitflips. However reality is too messy, so that the availability is not satisfactory enough to not use a human fallback, and then you are back to the assisted human error rate.
Current operation of a train is basically self-driving and supervising driver. So if you suggest "just use self-driving", you make the system less safe, not more, because it means just getting rid of the human supervisor. This is not currently done, because it would be too insecure. It might be done in the future, but rather because less safety is going to be accepted in favor for getting rid of the costly human.
A self-driving train will happily cause a crash, if the target is "drive with high-speed in that segment currently occupied by another train". The issue is that that becomes a target, not that there is a human in the loop.
Would you mind sharing some more of the why it is still an open problem? (I know I know little about this subject.) I understand satellite positioning isn't definitive and did not suggest it should be the /primary/ risk mitigating factor. But even as a tertiary or lower risk mitigating factor, this would have been a worthwhile investment. (Sorry for the bleak introduction of economics.)
And some problems are just _open_ to outside looks. My assistant tells me that the room for error in trains on one dimension ('left - right') is mere 10s of centimeters. I understand that problem. But what track you are on, isn't that obvious to the human in the cabin? (Again, simplifying, not flaming. Production ready safety is _hard_.) So could a vision based solution help count tracks from left to right? (Assistant: "US PTC, on about 54,000 route-miles, uses GPS for the lead locomotive plus crew-entered track information and switch position".) So you use the dimension that matters most in crashes (forwards / backwards) from satellite, augmented with vision or human based left-right positioning. Some trains already have sensors to sense things like leaves on the track to adjust speeding and braking... So the sensor based approach is already there.
Perhaps I am read as suggesting that the problem is easy, and that is not my intention. But for cars this was solved decades ago for routing and years ago for self-driving. I use that daily. Flying is organised with a very clear 6 sigma approach. Crashes are nearly always situations where events / risks stack. Trains (in Spain)...?
Tl;dr - Safety is a function of (human; process; technology) and there is clearly something lacking for trains, when compared to other types of transport.
So that's a lot of questions, that I all can't answer in full.
> Would you mind sharing some more of the why it is still an open problem?
Because for train sensors, there is a small drift which accumulates too much, hence it is synchronized by magnetic position markers in the tracks.
> I understand satellite positioning isn't definitive
Yes, and the point is it might be some solution if we would have the safety systems from 1860. But we do not and hence it needs to be more accurate than the current system, and it is not and in addition also more expensive (and subject to remote influence and reliance on potentially adversary nation states).
> Sorry for the bleak introduction of economics
No, economics is basically the primary constraint for safety. We could pay million people to make sure a single person does not stumble if we want total safety, but that's completely pointless, and also these million people would starve, because no one is producing food anymore.
> But even as a tertiary or lower risk mitigating factor
But for that it's too expensive and still unreliable. There is RCAS[1], which I already linked elsewhere in the thread. That's basically cutting edge and too expensive to introduce everywhere. Also previous safety systems are already not rolled out enough, which is causing these accidents, so rolling out yet another expensive system is not going to improve much.
> But what track you are on, isn't that obvious to the human in the cabin?
Yes, but if it is already known, why use GPS to calculate the same information with more error?
> So could a vision based solution help count tracks from left to right?
If the train driver or dispatcher makes use of vision to assess the train position, the error will be higher than the current sensor/electric-resistance based system.
> Assistant: "US PTC [...]
BTW, that seems to be a verbatim quote from https://www.railwayage.com/cs/the-way-forward/
> US PTC
According to Wikipedia, ASFA being introduced in 1975 is rolled out in the entire Spain network. ETCS which is the European standardization seems to be rolled out on all high-speed lines or currently being rolled out.
> So you use the dimension that matters most in crashes (forwards / backwards) from satellite, augmented with vision or human based left-right positioning.
Duh, that just sounds like trains would be operating blindly in space currently and need to assess their approximate position.
> Perhaps I am read as suggesting that the problem is easy, and that is not my intention.
No, the issue is that you are suggesting far more complicated solutions to already solved problems, with simpler systems in operation for decades. The current system is that nobody in the train even has a say on where the train goes, so the problem you are trying to prevent doesn't even exist. This IS why it's way more safer than a car. There are movement authorities, that define from where you are allowed (and have to) drive with your train. The problem is not, that the train driver isn't able to know where he is and hence drives wrong or too far (that would be prevented by the on-board safety system), but that the movement authority itself is faulty.
On an up-to-date signaling box (withing the past 6 decades), even the train dispatcher is no say in safety questions. You could put a monkey in charge of it. There would be chaos, but they would not cause a crash. The current job of a dispatcher in normal operation is to assign resources to prevent deadlocks and to implement vaguely defined priorities. (Basically the equivalent of the public thinking a programmer is a code monkey, while the actual point is clarifying stakeholder interests.)
The actual issue here is, that while these safety system have a very small false positive rate, they do have a too large false negative rate for there to be an acceptable availability. Hence in the latter case, the human needs to take over and then you are back to a much higher false positive rate. The issue would be alleviated by relaxing the constraints on cost per availability. Either by allowing for less availability or by reintroducing more humans, so that the system situation can be assessed by humans directly.
These accidents are caused as follows: the automatic system fails, because the internal state of the system does not match the actual state of the world. Hence the human needs to take over. But the human also needs to rely on some system to know the world state. Which systems can he trust, given that the whole situation is that the systems report that he can't trust them? So he needs to make some decision and in most cases, that decision is still correct. But there is still an error rate to that, and then something like what we are discussing happens.
> But for cars this was solved decades ago for routing and years ago for self-driving.
This is absolutely not solved for the cars, and to the sense that it is, it was solved and rolled out from the 1870s onward for railways using mechanical computing systems, the switch to using electricity happened in the 1910s.
> Flying is organised with a very clear 6 sigma approach.
You are correct, that aircraft safety systems need to be designed to a higher reliability than railways, because on a rail you can always stop in case of an error, while in the air such a choice doesn't lead to a safe state. But if you compare these safety systems, basically aircraft is operated like railways used to be operated in the 1920s. (My peers actually had a task to compare them and find ideas from one that can be applied to the other. While there is a lot of automation to be taken from rail that can be applied to the aerial industry, there is not really much that can be applied the other way.) It is also operated with the amount of personal than 1920s railways. That's because there is a lot of money in aircraft, while railways are mass-transit and need to be cheap. If you would operate today's railway with the amount of human oversight than in the past, it would be a lot safer than aircraft.
> there is clearly something lacking for trains, when compared to other types of transport.
What are you talking about? Railways are so much safer than cars. https://news.ycombinator.com/item?id=50288518 Yes, there are less accidents in planes than in trains, but this is due to what I already explained.
Thank you for your time and knowledge.
The current control system dates from the early 90s. They started rolling a modern system with more redundancies and automatic blocking in 2022 but it's still not fully deployed.
PZB is from 1934. Arguably always prioritizing deploying the most modern system on the highest profile lines, is what prevents us from properly rolling out the basic safety systems everywhere.
These high speed trains are pretty recent. Are we really not upgrading safety systems past the 1930's when upgrading a line from ~80kmph to ~300kmph?
I was only talking about the general problem domain, not about this specific location. The current safety system is also not was has caused this, likely. Or rather it has caused this by it's availability, not by it's safety.
Sorry, but ordering trains to drive manually after the system locks itself down, is the proper procedure. Of course you are still supposed to only send trains to free sections, but the error is in that step, not that they ordered a train to drive on it's own.
Also the fallback procedure is also aided by automatic systems and a sizable number of accidents in the past were caused by that information being faulty.
It is the proper procedure, and it does come (in a lot of countries, AFAIK) with a lot of guardrails:
There will be a so-called "train order" allowing passage of a specific section ignoring the safety system warnings, which has strict validity limits both in time and in the area where it applies, usually denoted by specific hectometric mileposts
When cab signaling is involved (LZB is one of these) and it malfunctions, the entire system normally changes to "driving on sight only" - which indeed limits the train to 40, 30, 20, sometimes - 15 km/h (see https://fr.wikipedia.org/wiki/Marche_%C3%A0_vue) or the line gets switched to lineside signaling - if it is available.
I don't know how you could violate all of those rules - both from the driver's side and from the traffic control side - but it sounds absolutely horrible. Like an override of at least 2 (if not 3) very explicit safety margins.
There are audio recording from this incident where you can hear the "train orders" being given and acknowledged.
They were the wrong orders apparently.
Why is that?
I remember a long time ago I was stuck in a train in a field a considerable time because of a similar failure and conductor just wouldn't move unless green
I was expecting the same here, conductor is the first to die so just refuse until the idiots fix their system
Many train systems have a trip-stop on signals that will automatically activate the trains's emergency brake if one passes a red signal.
It cannot be bypassed, if an operator must go through a red they have to manually reset the e-brake on each car once it passes the signal.
> Why is that?
What exactly?
A conductor always just gets a movement order. They are not able to assess whether the track is clear, that's the reason why the signalling system exists in the first place. How they get that movement order doesn't necessarily tell what the security steps were to back it.
Of course an attentive conductor can notice and mention there something being weird ("Doesn't the other train normally arrive first?"), but that's a rare lucky occasion, nothing to be relied on. And they just need to swallow the answer ("No, you go first today...") anyway.
> and conductor just wouldn't move unless green
The conductor won't ever move unless they get a movement authority, which most of the time is a green signal.
so in this case, couldnt conductor just say - make it green, not my problem, staying still until then
Why would they trust "a random blinking green light", but not a written and stamped order? Yes, they could do that (and sometimes it will actually prevent an accident), but it will be more being a snowflake, then being a decision for safety. A conductor is simply not responsible for whether it's safe to drive on a given track. Their equivalent is not a car driver.
The "random blinking green light" is generally more reliable
Being a snowflake here has 2 benefits:
1. Getting company managing tracks to fix the lights (they need to deal with angry customers and pay off damages for delays and more), so there's a direct path from unreliable lights to economic damage.
2. Avoid crashing and killing yourself and passengers on the offchance the person on the radio is a moron
> The "random blinking green light" is generally more reliable
Yes, but not because of the light itself, but due to the safety steps done before. Which should also have happened before the movement order in any other form.
As a train driver, you do not have the information required to distinguish these situations. If drivers simple refuse all movement orders, besides "green lights", running trains will become a rare occasion and the drivers will simply get fired for not doing their job. _The train driver is not the one who decides where a train goes._ For this safety-relevant decision a driver is (roughly speaking) nothing more than an innocent bystander.
You can think of a train driver more of the person who makes sure the automatic engine works correctly and also has an eye, that nothing weird happens. Most of your train driver education is knowing how the train works and how to get a broken train back working fast. In addition you also learn the safety protocols how you get the information in which state the engine should operate in. This is what train signals are, it's not really the equivalent of a green traffic light for car traffic.
Yes, but just with max. 40km/h
Depends. For example German Zs1 (https://en.wikipedia.org/wiki/Railway_signals_in_Germany#Zs1...) would give them the full clear outside of stations.
So that's why my train slowed to 40kmph when the Norwegian Telecom died for a few hours 3 weeks ago.
Yes, 40km/h is for driving on sight in most Europa. I think in France it's even 30km/h.
So based on this, the safest place to sit is presumably the middle carriage?
That has been true in the last century, but now the classic collisions are way rarer. See https://en.wikipedia.org/wiki/Eschede_train_disaster .
I would actually sit in front now. Sometimes you would be able to still get vital information and being actually able to act. But the safe decision would be no to enter the train.
That's where the bus hits you.
It's incredible the minister didn't resign after the January crash and I doubt he will now either. I understand that perhaps the problem started in the control centre but surely at the end of the day the minister is responsible - they don't get paid a fortune just to pass the blame on to the workers.
I fail to see how the two accidents, which seem to have very "down to earth" explanations (maintenance failure by workers, protocol override by operators) are being assumed to be a political responsibility so quickly by some and asking for the minister resignation. Now, two in a row seem to point to systemic issues. Is this responsibility of Adif or Renfe? maybe those state companies have to be thoroughly revised. is it due to lack of funding? maybe taxes should be raised, or money diverted from somewhere else (though seeing the spanish budget, it's hard to tell where from, both teachers and doctors are currently protesting for lack of funding). Maybe it is, in the end, a political directive that should have responsibilities declared in the executive. But seeing so many interested parties jumping to that conclusion literally hours after the crash means it's not done out of any reasoning but political animosity
> are being assumed to be a political responsibility so quickly by some and asking for the minister resignation
It sucks but that's how it is over here. It's "my" party the one in power when a catastrophe like this happens? It's not the politicians, it's someone else. Is it the "other" party? Surely the politicians are ultimately responsible.
This reflects on the news media as well.
> they don't get paid a fortune
Salary for ministers is around 80k euro per year. Yes significantly more than the minimum salary but hardly a fortune.
I can't imagine what every passenger who survived and the families of those who passed are going through. Such a catastrophic incident with multiple human failures is so tragic and so needless. From the controller to the driver, every single one of them should have said hey we shouldn't just ignore a signal and should investigate why it's on first. For this to be such a quick decision suggests it's not a one off. I would be extremely interested to see how many near misses or ignored signals there has been across Spain. The crash earlier this year was due to inexperienced workers and it seems as if significant training improvements and protocols are needed.
To have had this many incidents compared to other HSR networks around the world gives all the signs of a classic case of normalisation of deviance.
Sadly we'll probably watch a lot of individuals being blamed/scapegoated for systemic issues, but hopefully even Spanish politics can organise a proper investigation after this many deaths in a single year.
If as a driver you get told by the control centre to get the train moving again at maximum line speed, surely you should be able to trust that it is safe to do so?
I don't see how this would fall on the drivers at all.
My thoughts as well. It appears that such signal failures happen so often that the controllers regularly let trains run through red under verbal authorization. I don't know the exact protocols, but I would expect a driver to ask for confirmation, as happened here, and then trust the control centre. However, there is clearly a huge systemic problem, with added complexity for the controllers, and there should at least be double checks in the control centre.
The person who controls whether the light shows red or green and gives verbal authorization is the same. It's both a tool, to transmit the same information.
Trust but verify. You wouldn't drive your car into a lake just because your satnav tells you to, would you?
As we also know from the aviation industry, controllers are human and therefore will make mistakes. A controller shouldn't issue a wrong order, but a driver also shouldn't accept an order they aren't certain about, and there should be technical systems backing both sides of that up.
It doesn't "fall" on the driver, but it doesn't "fall" solely on the controller either. It's a systemic issue, and many pieces have to fail for a disaster to happen.
> As we also know from the aviation industry
The aviation industry operates like the railway industry ~1920 (somewhat by necessity).
> You wouldn't drive your car into a lake just because your satnav tells you to, would you?
You are assuming a train driver is equivalent to a car driver. He's not, the analog would be more a repairmen that you always take with you on journeys.
> From the controller to the driver, every single one of them should have said hey we shouldn't just ignore a signal
A train driver has zero insight in what is going on and where other trains are. They do need to trust the orders they get, if you want to question every order you will never start to drive.
The train driver's job is to say: "it's red, I'm no moving, not my problem, you make it green if you want me to go"
That's why it's a union job
Also he's the first person to die on accident so it's that and the safety of everyone else
They do not 'need' to follow others, they just need to refuse unless all safe and green.
If they never start to drive it's the companies problem to deal with angry non-dead customers
That is not how the procedures work, so the driver would be fired as soon as they did that.
Every train network has situations where drivers are authorised to proceed against a red signal. Not usually at full speed though.
They wouldn't get fired. They are protected both in the US and Spain.
At most face unlikely discipline.
Here they would even have valid reason to refuse - order was given for full speed, regulation says low speed ready to stop.
So a "No, i'm not doing that, you're moron" would be justified.
No, he would get fired, because that's not his job. It would be like the machine engineer on a race track refusing to fix the ignition, because somewhere on the track he doesn't see, the asphalt might be broken.
So it’s clearly an industry wide problem then. The more I read the more thankful I am California can’t build high speed rail.
Except that car vs. rail deaths per 100 million person km is 0.7 vs. 0.035 and per 100 million person hours 25 vs. 2 (https://etsc.eu/transport-safety-performance-in-the-eu-a-sta... or https://injuryfacts.nsc.org/home-and-community/safety-topics...). It's more of an human wide problem, rather than industry wide problem. If you want to be safe, you need stay at home, but that's also not going to be actually safe.
The point of this discussion is that people blame the train driver or say "use self-driving", when that is not going to solve the problem, because that's not where the error is. The error is not there, *because* for rail making errors operating the train is not going to cause an accident (most of the time), which is not the case for cars.
The driver always has the final say on whether the train is safe to move. They are not just a button and lever pusher. They have to analyse whether the decision is correct. Hence why I stated there were multiple points of failure. The controller and the driver both went with full speed despite it being completely against regulation on a red signal.
> They are not just a button and lever pusher.
I agree. But the decision whether all switches are locked in the correct position and that the track is clear, is not one that he is supposed to make and he also doesn't have the information for it.
> The controller and the driver both went with full speed despite it being completely against regulation on a red signal.
Going full-speed in the case of a red signal is not necessarily against regulation. Also the information of "red signal" or "you can go" is both a signal from the same source. There is no information for the driver to make a different decision over the dispatcher.
I can't comment about whether the current accident was caused by going against regulation, and more important by whom and against which. We need to wait for the proper safety report, not for a BBC article. If you take the latter at face value, it would mean there are clowns operating, which I am sure, is not actually the case. The real error is always way more subtle, otherwise it would not have happened.
> The train driver's job is to say: "it's red, I'm no moving, not my problem, you make it green if you want me to go"
A written movement authority is a signal as well in railway terms. It's more like "I don't trust your second traffic light, that you placed next to the normal one, because the normal one is broken".
The even more insane thing to me is that the controller knew that there was another train in that block, knowing it was no signal failure. The driving train should’ve never been commanded to continue on full speed by the controller.
Trains shouldn’t even be manually driven in 2026. We have self driving cars now which is a way harder problem to solve. Relying on a system of signals and driver attentiveness is archaic.
> Trains shouldn’t even be manually driven in 2026
Potentially, but the point is that signallers should physically be unable to allow trains to proceed, and drivers should physically not be able to override them. British signalling from the 1800s has physical interlocks; this isn't a matter of manual versus automatic train operation. It is clearly systemic.
How do you pass a defective signal then? How do you couple a second locomotive? How do you add cars?
We do not have self-driving cars in all environments that clear the necessary security regulation! Also the socially accepted number of deaths due to trains is way below that for cars. The current accident is just a Tuesday for car traffic. (Spain recorded 1785 road deaths in 2024[1], that's 4.9 per day) Detecting objects (and future objects!) on a train track is still a hard problem.
The current system is basically that the train is driven both automatically and by a professional attentive driver at the same time. Changing it to just the automatic system is going to reduce safety. Also the real world is way to messy to reach acceptable availability rates with only automatic systems.
[1] https://www.itf-oecd.org/sites/default/files/spain-road-safe...
> Detecting objects (and future objects!) on a train track is still a hard problem.
In this case the “object” was another train. That’s not the “hard” part of the problem. This sounds like an apologist comment.
What difference does that make, genuinely?
Are there systems that can warn drivers and prevent collisions without external signaling, like ACAS?
Yeah, that's a solved problem. Did you even read my whole comment? I answered why "just use the system from self-driving cars" wouldn't work and in fact would be less safe than the current system.
Driving a train and driving a car are two completely different challenges.
If signaling worked properly then it wouldn't matter whether the train was driven autonomously or not, the "problem" becomes trivial. Establishing and continuously verifying that signaling is working properly is the hard part.
Without signaling you're limited to crawling speed just like humans are when they don't have any signalling .