NY Rail Accident

Here we are talking about computer controlled cars, and if there is anything that should be almost 100% computer controlled....it should be trains.

London's Docklands Light Railway is already.

However, no automated system will cater for an obstruction on the track that remains undetected by the signalling system. That's why drivers are put on the front; it's the driver's life that is the first one to be threatened should something inappropriate occur.

There is usually a lot more to these sort of things that we average public news readers get so I am going to play a bit of devils advocate on this topic.

First off from my and most peoples experiences of falling asleep on the job how many have did this after going to a different shift even though you thought you never would? I have and I have done it a number of times.

Then from that how many ended up on a different shift because you had a boss or manger put you there only because it suited his whim/office politics and your being moved had nothing to do with anything else? For me shift jumping was usually due to poor management of office politics. Rarely did I ever willingly choose to jump from one shift to another.

My point is a person falling asleep at their job has a number of other factors in it that should have been addressed other than 'How can we make the machines smarter?' of which did this person have to change shifts because he wanted to or was he forced to and if forced to why and by whom?

I for one feel bad for the guy who was running the train. So far it has been reported that he recently changed shifts and that had a lot to do with is having fallen asleep.

Now my question is why did he change shifts? Did he request it or was he forced into it against his own better judgement and this is the end result?

Be that as it may... the simple truth is... in this day and age, why that train control system did not know what the maximum speed was for that segment of track and was able to limit its speed to that maximum is... beyond me.

We are far past the time period where we "needed" human operators for such equipment. Their function should be to monitor the operating systems... not operating those same systems.

There is no excuse for this kind of accident in 2013.

There is indeed an excuse for this accident in 2013. Or a reason, at least. Money! Greed. Avarice.

The technology exists, and has for some time. It costs money to implement it and the railroads don't want to spend it. Systems such as this have been mandated by government for years, but railroad management has been $ucce$$ful in delaying implementation to save money.

This fact has been reported in the news media.

Agreed...

Sad... but true.

this lack of implementation quite often influenced by unions claiming to be protecting jobs that would be taken by machines a redundant computer control system doesn't text while driving

Nor fall asleep.

Implementing this system would not cut jobs. Intead of driving the train... these emplyees would be overseeing the driving done by the control system.

Featherbedding.

Not familar with the term... explain please. Though I can imagine what ya meant.

Term comes from the fact the firemen and brakemen were still riding trains when they were not needed. Diesel locomotives and air brakes made them obsolete, but unions prevailed for a long time and they kept their jobs.

Someone more intimate with the term may want to correct me.

Featherbedding

Computer may not fall asleep but they sure seem to reboot or want to update quite often.

Obviously... we are not talking about a MS Windows equipped PC, more like a PLC or other dedicated mission specific controller.

Or they would be taking extended naps instead of 'nodding off'. So, move the control to a control room where one operator can operate 100 trains then see what happens when he goes to sleep. -- JHF

Some of what you point out is the reason manufacturing was shipped oversea. It was easier to do that then to fight with unions over automation and what to do with those employees it would displace.

It also seems that every time the US Post Office raises the cost of letter mail the media puts it in print like they are robbing the poor. Usually by the time they are allowed the increase which takes an act of congress. They have operated in the red off tax dollars for a few years. So I wonder if they didn't have this automation how much first class postage would cost. It would take a lot of clerks to file by hand what one of those machines can do. Any reduction in force by the post office was done with early retirement bonuses no body is usually laid off.

Yep. I wasn't looking to derail the thread, , but Fredski had a good point. I believe the railroad workers still have a pretty strong union, and they will fight any type of automation. that replaces people.

If there is regulation to provide safety for such segments of tracks then how come trains without such provisions are allowed to run?. Safety dept. of Govt. should take immediate action or such accidents can repeat again and loss of innocent lives will continue.

generally speaking I agree with you. I speculate that if commercial airliners were 100% controlled by on-board computers (its heading that way) and two planes were leaving from LA to NY with one being piloted by a traditional-pilot and the other fully automated and passengers were given a choice they'd opt for a human at the controls.

Agreed...

But... there is a huge difference between an aircraft and a train. We are talking about maximum speeds, who would argue with that?

with the proper sensors monitoring the tracks themselves as well as the engine those speeds would have never been seen, no question

And over 80% of passengers would prefer a male pilot.

Ever hear of the phrase "featherbedding" I think it was coined because of railroad unions.

The news report I saw today stated that there was a deadman plate the operator was standing on but it had not been determined if it was fully operational.

Automation is not impossible, but it is expensive. To retrofit the entire industry would cost millions if not more. That cost compared to the value of one life? I am glad that is not my responsibility to decide.

I read about an automatic system built in California, the system had not been fully tested before construction began. One of the Engineers in charge of the program was listed in my textbook as a whistleblower because after attempting to inform his superiors about the lack of testing he went to the government with his concerns. The project went ahead and the Engineer was punished. From operational reports it sounds like they tested the system as it went live and there were some mishaps but I don't recall any fatalities (but that was luck).

A lower cost system might be to install a computer with a camera that watches for sleepy drivers, I heard there is a system for cars that does the same thing. It could be connected to the deadman system or an alarm that might wake the drowsy operator.

Drew K

How about existing technology that would not be expensive to implement.

Dead man's switch - Wikipedia, the free encyclopedia
The foot pedal seems childishly simple to disable.

I don't understand why automation for trains would be so expensive?

Wouldn't it be a much simpler version of what's used in planes?

It's just an engine sitting on a track. No turning or anything.

Seems like they would just use the existing GPS system, program in specified speeds for varying sections of rail, adjust for loads, emergency braking for obstructions, etc....

I'm with No60...it's dumb to use human control on a thing that spends most of it's time simply moving forward on a track.

There are appx. 25,000 diesel locomotives in service in the USA today.

A cost estimate of $500.00 USD each to retrofit gives a very conservative cost of $12,5000,000.00 USD.

It's all about the money!

How much are the lawsuits going to cost?

On advice from my attorney, I cannot answer that question.

It's a calculated risk. Up till now, it has paid off.

This has been a really bad year for Metro-North.

Lyn,

Correct fig would be $12,500,000/-. Not a big amount compared to loss of human life.Railroad companies can recover the cost by increasing their fare moderately on long term basis.

Whatever.

The "real" figure seems to be $900,000,000.00 and the railroads want US taxpayers to pay for it, yet they still don't want to do it.

This link is to a case study of the automated system I wrote about above. I do not know how much the system cost and it seems to have worked out in the end but was not without its problems early on.

As for why it has not been implemented elsewhere since...I don't know.

Drew K

Not necessarily so... my $150 GPS unit knows the maximum allowed speed on the roads I drive on. The GPS system manufacturer could easily include tracks.

As to modifying the train control systems, someone said there was less then 30,000 locomotives in use in the USA, 30,000 units at $1,000 a pop (just to modify the control systems) is only $30 million.

There is more than just the cost of a life or two to consider here. The 30 mill seems cheap to me in comparison to the dollars involved in the loss incurred from the loss of the rolling stock, the track systems and the impact on the human side of the equation.

Issue a GPS unit to each locomotive engineer... he plugs it in to his cab when he boards and Bob is your Uncle.

He might have some mild form of epilepsy or narcolepsy , but according to privacy laws, that can't be divulged.....

About everyone at sometime while driving has had a spell where they start to doze off.

Driving long distances on the highway that broken white line's hypnotic effect can lull one to doze. Most of us catch ourselves and do something about it. They never loose complete control. So a dead man switch on a train just doesn't work in this case.

Some don't get so lucky and have an accident. In this case the train engineer luck wasn't with him.

Most freight trains have locator's on them. It's that light at the back that replaced the caboose. It locates the end of the train for the engineer. So he knows what he has cleared back there. There are a lot of old single track bridges. That he has to know he's cleared for traffic coming the opposite direction. There is also systems monitoring it how else is he to know something coming the other direction to let it get across before he gets to the single rail bridge.

Whether the passenger trains use this technology or not I don't know. They may. But is anything better then having a person at the controls. Even machines fail. I do not think at this time they have ruled some that out.

My suggestion doesn't cost much more money or any at all. A little bit of chatter between the engineer and persons monitoring they systems in place. To distract them from things that would cause one to doze. Even if it's done on a station to station level.

Heck yeah!

Ever been on a lonely stretch of highway at night, and had the cruise control on?

Even if you're rested up, it's hard to stay awake.

I couldn't imagine operating a train, with mile after mile of nothing really to do.

Or just put an alarm on the control panel...that wakes up the operator and lets him know that the train is on auto-slowdown.

For the long boring stretches...why not let the guy catch a few winks?

asleep at the wheel, great idea

You CAN'T be serious! There is no auto-slowdown.

They should electrify the seat and rig it to shock the engineer if he doesn't push the big red button when it lights up, or just have that stop the train.

this might get their attention

My point was....that slowing down for corners, and approaching stations, (which are the same every day), should be automatic, and not dependent on a human.

Computerizing repetitive tasks should not be hard or expensive, and would not replace the operator. It would be just like auto-pilot on a plane...but much, much simpler.

ever spend any time in the Atlanta airport?

Tell it to the railroads: Metro-North Crash Seen Thwarting Delay of Automatic Braking

A telling tidbit from the article, "Metro-North and other railroads had lobbied for the delay."

Hey, only 4 people were killed this time, what's all the fuss?

This from::Passenger who survived New York train crash intends to sue railroad

"MTA spokesman Salvatore Arena said the agency began work to install positive train control on the Long Island Rail Road and Metro-North Railroad in 2009, budgeting nearly $600 million for its installation, including $428 million last month for a system integrator. The cost for full implementation is estimated at $900 million."

I'd be trying to find something much more reasonable in cost.

Your first link sums it all up.

Overthink

Lobbying

Politics

Money

Seems to me that they could just use a simplified version of the systems that run roller coasters.

PHOTOS: Metro-North engineer minutes after crash

Thanks all for your worthy comments. We as engineers would always search for solutions to tragic accidents so that they never happen again. I do not know if our our voices are heard by any one in authority, but hope they would reach them soon.

Yeah right.

How many times have you actually seen a suggestion box get opened?

If they do open them, it's to identify trouble makers that don't want to go with the flow.

We use a monitoring system that centers a camera focus on the operator eyes and head.

The software maintains an acceptable pixel zone that the operator must keep their head within and they are required to look at the camera often.

If the eyes start to close or the head does not stay within the zone, a loud audible alarm is automatically triggered.

The system is monitored remotely and if the alarm triggers three times, the equipment is automatically shutdown and the operator removed & replaced.

Many operators do not like the system but our accident rate caused by sleeping has been decreased by over 98%.

That's too cheap and easy.

We need a system manufacturer that can afford to finance a couple of political campaigns. That's the system that will be implemented by law.

The government has mandated that a computer control system be installed on all freight and passenger trains by 2015. That would monitor speed and location and would over ride the engineer.

This commuter train company has already started the installation.

http://q13fox.com/2013/12/04/new-technologies-aimed-at-preventing-passenger-train-accidents/#axzz2mcGLldW4

And these systems need not be terribly expensive either. My car's GPS seems to know when the speed limit changes to within a few feet of the sign posted on the road. What will happen, is that the safety folks will try to put so many bells and whistles on it to take care of every conceivable situation, it will delay its implementation and make it cost exponentially more than a simpler system would.

I was thinking about that when I woke up this morning, and for some reason, was thinking about this thread.

They could just use something like those Tom-Tom units you see advertised on TV.

"Slow Down Now"

"Of the 996 cab cars used to drive Metro-North's fleet, 331 are equipped with what's called a "dead man's switch," a pedal the operator must keep pressed to keep the train moving. The remaining 665 cab cars in Metro-North's fleet have such "alerter" systems. With that system, if the engineer doesn't respond to beeping signals at certain intervals, the train's brakes are activated."

http://www.usatoday.com/story/news/nation/2013/12/05/cab-car-in-ny-derailment-lacked-alert-system/3886757/

I learn from my Australian friend that they have "Vigilance control system" installed in the engines of locomotives. As per this system there is hooter inside the driver's cabin which hoots periodically and driver has to stop it, in case it is not stopped by driver, if he is napping, then engine is automatically stopped. This is simple device which railroad companies could develop locally or buy it. It could save many lives.

Most of that is covered above. Dead Man's handle covers driver collapse but not those brief sleeps that all of us take when tired or bored.

UK is covered by a system that alarms at every signal and must be cancelled by the driver before an automatic emergency stop. It is not in use in america much because of distances involved, and will soon be outdated by GPS systems as mentioned above. A GPS system would apply well to Indian and other continental rail routes.

Not many trains have guards any more, not sure about that network. Most are re-deployed as Passenger Service with snack trolleys which does not apply to a local commuter service. Too distracted then anyway to provide a monitoring service.

I do not know how GPS system will wake up the sleepy driver?.

The GPS would be the Core of a system providing location and linking that to detailed track mapping. Things like curves, or just speed limits, can be readily programmed in, checked against actual speed and 'bad results' being wired out to the trains brake.