Driverless trains.........

[Diesel-Electric]

For some reason, I find this story depressing:

http://ozhouse.org/2012/02/20/austr...ork-of-driverless-freight-trains-job-wipeout/

http://www.rawstory.com/rs/2012/02/...tm_campaign=Feed:+TheRawStory+(The+Raw+Story)


Now. I'm not anti-technology(I've benefited a lot from it) and I know driverless trains are tried and proven with excellent safety, but you know, even though a human being can make disastrous mistakes, I still would always like to have a human being at the controls.

I feel that technology should have it's limits!

Just me opinion!

 

[LoudMusic]

I am very pro-automation. I am also very pro "a human needs to keep an eye on this stuff!". I see no reason to remove the humans from the trains, but automating the mechanics is a good idea. It can provide more efficient time tables, and more efficient fuel consumption. It will also reduce disasters. All of these things together will increase respect for trains, and in turn increase the use of trains. And more trains == more better ;D

It just requires a pretty significant infrastructure to create a reliable automated system. There might be a reduction in the jobs available driving the trains, but there will be more jobs created maintaining the systems which automate the trains. And again, if it's successful, there will be more trains.

Thank you for sharing the article.

 

[LoudMusic]

From the article, "... annual net earnings of US$5.8 billion ..."

With that kind of profit they're not automating to eliminate job costs. They're automating for all the genuinely positive reasons.

“Our business, no matter what, even with the autonomous train programme will still need more employees in five years time than we need today in total,”

As I said in my previous post - they're just transitioning from on-train employees to other positions which maintain the automation and operate trains remotely. An engineer doesn't have to spend a week on a train away from his family anymore. He can work an 8 hour shift, and go home to his wife and kids while the next guy sits at the office and keeps an eye on the train.

 

[TrinityJayOne]

Despite these trains being on the other side of the country I can shed a little more light on this. As far as I know the trains are being driven remotely rather than having staff in the cab, which is something they've already tested and proven to work. Also, the mining industry in Australia is highly profitable and employees make very good money, the downside being that they are quite literally working in the middle of nowhere in one of the hottest parts of Australia. I think automation will work well for them because of the conditions and because of the relative safety. These trains aren't navigating a busy suburban network, it's a train line built by the mining company solely to run their trains on and probably nothing but straight track for miles & miles & miles.

 

[Cjcrescent]

Despite these trains being on the other side of the country I can shed a little more light on this. As far as I know the trains are being driven remotely rather than having staff in the cab, which is something they've already tested and proven to work. Also, the mining industry in Australia is highly profitable and employees make very good money, the downside being that they are quite literally working in the middle of nowhere in one of the hottest parts of Australia. I think automation will work well for them because of the conditions and because of the relative safety. These trains aren't navigating a busy suburban network, it's a train line built by the mining company solely to run their trains on and probably nothing but straight track for miles & miles & miles.

Yeah, while it has been proven in that situation, I don't think it would work on this side of the big pond. There's too many grade crossings, towns, cities, curves and never mind the forests, etc. that are close to the ROW, just here in Alabama.

During the blizzard here of 1993, there was a situation where several large trees had fallen down onto the tracks from the weight of snow and ice. A train barely missed derailing because of visibility problems, the engineers couldn't see the trees down across the tracks. While they were only moving @10mph, and just thought it was a snow drift, they still hit them with enough force to do about $50,000 worth of damage to the front of the lead locomotive. I did see some pictures of the loco after they cleared the tracks and got the loco back here to Birmingham. The front end looked totally wrecked.

They were able to stop the train without derailing. Imagine how bad the wreck would have been if it was under remote control, and moving at a faster rate of speed. Radar, night vision, and cameras, even humans in the cabs, wouldn't have detected the obstruction soon enough to stop safely.

 

[LoudMusic]

Yeah, while it has been proven in that situation, I don't think it would work on this side of the big pond. There's too many grade crossings, towns, cities, curves and never mind the forests, etc. that are close to the ROW, just here in Alabama.

During the blizzard here of 1993, there was a situation where several large trees had fallen down onto the tracks from the weight of snow and ice. A train barely missed derailing because of visibility problems, the engineers couldn't see the trees down across the tracks. While they were only moving @10mph, and just thought it was a snow drift, they still hit them with enough force to do about $50,000 worth of damage to the front of the lead locomotive. I did see some pictures of the loco after they cleared the tracks and got the loco back here to Birmingham. The front end looked totally wrecked.

They were able to stop the train without derailing. Imagine how bad the wreck would have been if it was under remote control, and moving at a faster rate of speed. Radar, night vision, and cameras, even humans in the cabs, wouldn't have detected the obstruction soon enough to stop safely.

I think your assumptions about what's possible with sensors and computer processing are rather antiquated. Here are some fun videos.

ABC News on Google Car
http://www.youtube.com/watch?v=X0I5DHOETFE

Some Google Car test drive
http://www.youtube.com/watch?v=-cj375UZyaI

Quad-rotor aircraft autonomously building structure
http://www.youtube.com/watch?v=oFtP3y8Rf9g

Quad-rotor aircraft autonomously building structure in a group
http://www.youtube.com/watch?v=TI760jcFV2s

Aircraft circling each other and adapting to each others' position
http://www.youtube.com/watch?v=tdZAIt7OrPs

Autonomous quad-rotors compensating for hardware failure
http://www.youtube.com/watch?v=-cZv5oKABPQ

Jeremy Clarkson in a BMW 3 series
http://www.youtube.com/watch?v=DRF_KaWzxq4

This one is long but probably the most educational. It talks about how the Google Car works and handles real world conditions, adapting and compensating for changes in the environment it is attempting to navigate.
http://www.youtube.com/watch?v=YXylqtEQ0tk

So these are all examples of autonomous operation of vehicles which require significantly more control and environment input than trains need and they seem to be doing quite well. That's not to say it's perfect or that I'd be willing to hop in an autonomous car and go for a ride just yet, but I don't think we're very far away.

Heck, they're even doing cool stuff with Lego Mindstorm. Though the results are significantly less impressive

 

[D&J RailRoad]

I just don't think computer control is that perfect yet to allow the tons of frieght to go unattended.

 

[LoudMusic]

I just don't think computer control is that perfect yet to allow the tons of frieght to go unattended.

Ah yes, that's the key. It needs to be put into use as a backup operator to a team of humans. Configure it with certain timeouts and alert systems before it finally just makes the command decisions for the locomotive.

When the system observes a hold signal it would start a timer. After a certain amount of time (half a minute perhaps) it would turn on a red light or bing or something. If the engineer still hasn't taken action it could alert an off-site 'remote control' engineer, "Hey Bob is asleep and I'm supposed to be stopping!" The remote engineer can either hit the breaks or radio the engineer, "Bob, wake up!". If after more time has passed the train still does not slow down the automated system could apply breaks itself.

There could also be an acknowledgement system by the engineer. The automated system determines a control change is necessary and presents the engineer with the change options. "That signal said I need to stop. Do you confirm?" and start a count-down timer. The engineer could discard the command, perform it him/her self, or just let the automated system handle it when the timer runs out.

"The train has cleared the slow order and would like to resume normal speed. Please confirm." Same situation. Let it go and the system will speed up the train when the timer runs out.

Eventually having all the trains talk to each other could result in some profound efficiency improvements. One train routed for a siding could calculate an opportunity to run slower, even coasting from miles out, approaching the siding and burning less fuel and still arrive before the through train arrives. The through train could be told that due to the the fuel saving on the approaching train it can run a little faster in order to arrive at its destination a little earlier. Multiply that across thousands of trains and everything starts arriving at destinations minutes earlier, and thus shipping out minutes earlier, and more trains get run in a day.

Computer control can also be more efficient with throttle control over long hauls through varying grades if it has detailed information about the route. That information would build up over time in a database being updated by all the trains that ran through the area. Then they could pull from that information to make small adjustments to throttle control squeaking out minor amounts of improved economy. The huge amount of fuel burned by trains makes every little bit count.