Considering the advancement of technology, what are the things they do that can’t be replaced by automation?
In: 0
Brake because they see something ahead which wouldn’t be picked up by an autopilot.
Slow or stop because they hear on the radio there’s an issue ahead, without having to rely on data transmission in a bad signal area.
Also to add, the technology might theoretically exist to replicate some of this, but it’s not even close to cost effective to roll it out in the kind of total and error proof way that would be needed.
Brake because they see something ahead which wouldn’t be picked up by an autopilot.
Slow or stop because they hear on the radio there’s an issue ahead, without having to rely on data transmission in a bad signal area.
Also to add, the technology might theoretically exist to replicate some of this, but it’s not even close to cost effective to roll it out in the kind of total and error proof way that would be needed.
It takes time for the new technology to be deployed on a large scale. It takes time and resources to build all the sensors into the infrastructure and ensure everything talks with each other and works properly. Most train equipment have a service life of 40 years and the tracks have a service life of 100 years so it will take some time to upgrade everything so it can become fully autonomous. Currently I only know of one fully autonomous rail line and that is a newly built line through London.
So train drivers have to make sure that the train does not run a red signal, that they do not hit someone or something on the tracks, that the doors do not close on people, etc. And that is also assuming the train have auto throttle which most do not yet have. But even if you manage to install sensors and fences and such so that the trains can drive themselves without hurting anyone you still need someone on the train to handle unusual events. For example people or debris blocking the sensors, sensors or signalling equipment failing, driving the train in case of emergencies, shifting the train in the yard, hooking up to dead trains to tow them on, etc. There are lots of things that still require someone on the train.
Brake because they see something ahead which wouldn’t be picked up by an autopilot.
Slow or stop because they hear on the radio there’s an issue ahead, without having to rely on data transmission in a bad signal area.
Also to add, the technology might theoretically exist to replicate some of this, but it’s not even close to cost effective to roll it out in the kind of total and error proof way that would be needed.
1 Scanning for any unforeseen event.
Along the route: Tree branch fell on the rail, people/animal/vehicle incursions, sabotage.
On the platform: check no one is being silly while the train approach. Brake to prevent people to die.
2 On emergency: marshaling the evacuation.
3 All the time: provide a point of contact for the train, for whatever authority or entity needs to contact the train, and vice versa. Reset or fix any basic mechanical issue, including for example, a passenger operating the emergency brake for no reason.
4 Most importantly: take decisions. Decision making is very very important.
All of the above can be automated, but for now the cost and reliability is not within range of having a real person there with appropriate training.
It takes time for the new technology to be deployed on a large scale. It takes time and resources to build all the sensors into the infrastructure and ensure everything talks with each other and works properly. Most train equipment have a service life of 40 years and the tracks have a service life of 100 years so it will take some time to upgrade everything so it can become fully autonomous. Currently I only know of one fully autonomous rail line and that is a newly built line through London.
So train drivers have to make sure that the train does not run a red signal, that they do not hit someone or something on the tracks, that the doors do not close on people, etc. And that is also assuming the train have auto throttle which most do not yet have. But even if you manage to install sensors and fences and such so that the trains can drive themselves without hurting anyone you still need someone on the train to handle unusual events. For example people or debris blocking the sensors, sensors or signalling equipment failing, driving the train in case of emergencies, shifting the train in the yard, hooking up to dead trains to tow them on, etc. There are lots of things that still require someone on the train.
It takes time for the new technology to be deployed on a large scale. It takes time and resources to build all the sensors into the infrastructure and ensure everything talks with each other and works properly. Most train equipment have a service life of 40 years and the tracks have a service life of 100 years so it will take some time to upgrade everything so it can become fully autonomous. Currently I only know of one fully autonomous rail line and that is a newly built line through London.
So train drivers have to make sure that the train does not run a red signal, that they do not hit someone or something on the tracks, that the doors do not close on people, etc. And that is also assuming the train have auto throttle which most do not yet have. But even if you manage to install sensors and fences and such so that the trains can drive themselves without hurting anyone you still need someone on the train to handle unusual events. For example people or debris blocking the sensors, sensors or signalling equipment failing, driving the train in case of emergencies, shifting the train in the yard, hooking up to dead trains to tow them on, etc. There are lots of things that still require someone on the train.
As someone who works with camera vision systems a lot, they are not nearly as good as human eyes. Differing light conditions, unusually shaped objects, people with suspicious postures.. all of that is really hard to compute around.
Machine Vision works best in really repeatable conditions with things the machine has seen before. Humans are highly adaptable to weird things happening.
So in reality we use both, we bolt a camera underneath the train with a specific light to shine on the tracks looking for damage to the metal, but we use a driver to watch out for unusual weather and suicide attempts
1 Scanning for any unforeseen event.
Along the route: Tree branch fell on the rail, people/animal/vehicle incursions, sabotage.
On the platform: check no one is being silly while the train approach. Brake to prevent people to die.
2 On emergency: marshaling the evacuation.
3 All the time: provide a point of contact for the train, for whatever authority or entity needs to contact the train, and vice versa. Reset or fix any basic mechanical issue, including for example, a passenger operating the emergency brake for no reason.
4 Most importantly: take decisions. Decision making is very very important.
All of the above can be automated, but for now the cost and reliability is not within range of having a real person there with appropriate training.
1 Scanning for any unforeseen event.
Along the route: Tree branch fell on the rail, people/animal/vehicle incursions, sabotage.
On the platform: check no one is being silly while the train approach. Brake to prevent people to die.
2 On emergency: marshaling the evacuation.
3 All the time: provide a point of contact for the train, for whatever authority or entity needs to contact the train, and vice versa. Reset or fix any basic mechanical issue, including for example, a passenger operating the emergency brake for no reason.
4 Most importantly: take decisions. Decision making is very very important.
All of the above can be automated, but for now the cost and reliability is not within range of having a real person there with appropriate training.
Latest Answers