How about the failure or other inability of the signalling system to issue a movement authority? If the interlocking won't allow a route to be set (and thus can't issue a movement authority), it's likely that you'd still be reliant on a driver acting on verbal instructions - proceeding with caution, checking that the route's set correctly and so on. That sort of thing happens fairly regularly, for all sorts of reasons, and would still require a fully trained driver - route knowledge and thorough rules knowledge - to deal with it.
Examining the line - again, requires thorough route knowledge (where exactly to examine?).
Even a very low speed move into the nearest station (how far might that be?) isn't that simple. How does the driver know which signals (or block markers, or whatever we'd end up with) apply to him, and how far his authority to move can take him? Speeds through crossovers? Other local instructions?
It may be possible for the train captain to drive by sight and be prepared to stop short of any obstruction at low speeds eg 15mph max. This has already been permitted on the Thameslink inner core and east london line under certain signal failure conditions.
We already have driverless trains at 80kph on the Docklands light railway which is now quite a complex network. I would imagine that it would happen on intercity routes last if at all and initially on fairly self contained lines like merseyrail or Thameslink inner core and between St Pancras and Bedford.
Also, when a train was being driven manually under auto driver failure conditions, a control supervisor could montor the route ahead via cameras mounted on the front of the train and be in constant radio contact with the train captain acting as a "second man"
I don't think it is likely to happen on any wide scale in a timeframe that threatens many current drivers jobs, which I suspect is the principal unstated fear.
--- old post above --- --- new post below ---
Whilst I am sure you know best, I wonder if the complexities of the operational environment within which the "simple.... guided vehicle..." runs may have escaped you. It is not quite as simple as a car. Tomnick and O L Leigh have suggested pertinent reasons why it is more complex than that.
I'm not convinced, trains do not have to share a road with pedestrians, bicyles and 40 tonne artics all at the same time and cope with constant unprofessional and downright dangerous driving by assorted numpties.
I will say again that ECTS does not envisage a driverless train in the open environment.
Indeed, but it is the logical next step, and has already happened on the now quite complex 50mph Docklands Light Railway. As above I would imagine fairly self contained lines with frequent services like Merseyrail as the first to see train captains replace drivers. I would imagine a prerequsite would be full grade separation, ie closure of all pedestrian, vehicle and occupation level crossings and possibly platform edge doors, with high speed mixed traffic lines like the WCML many years off from consideration if ever.
The economics also favour short lines with frequent services as they have lots of trains (so lots of drivers) and comparatively low fares. Long distance has fewer trains with much higher fares so the drivers are a much smaller part of the overall cost and the longer distance means installing the technology along the track costs far more.
As above I think some of the "it can't be done" can be translated "it mustn't be done as it would put me out of a job" which is not a realistic fear for the vast majority of current drivers.
One thing that might threaten drivers jobs would be if the road network went over to driverless vehicles and the rail network didn't; because the road network will inevitably go over from people owning vehicles to people summoning driverless vehicles on a mobile phone app at short notice, at a fraction of the rail fare.
This will decimate rail travel unless rail costs can be heavily cut, especially as if all road vehicles were driverless a significant increase in permitted speeds on motorways and high quality dual carriageways would be likely; with vehicles formed into rolling convoys ie virtual trains, joining and splitting as needed on the move. Such automated roads/motorways would also have massively more capacity than current roads (in the same way that moving block signalling gives railways more capacity) and could result in a rail network that makes the serpell report look generous in 20-30 years time.
Be careful what you wish for
--- old post above --- --- new post below ---
I'm sure I dealt with this point adequately well in a previous thread some months back.
Click.
A driverless car only has to recognise when there is something in it's path in order to take evasive action. A driverless train will not be able to take avoiding action and, in most cases,
will hit whatever the obstruction is. Therefore it needs to know exactly what it's hit in order to formulate it's next response.
Whatever you believe about the levels of advancement in computer technology the human brain is incredibly good at quickly identifying objects. We can easily tell the difference between a person and, say, a deer because we have the experience and ability to discern between the entire subset of "person" objects. Therefore we can tell that we're looking at a person whether they are large or small, standing, crouching, running, partially obscured, etc, etc, etc. We can also apply judgement based on experience. A person lurking at the side of the track could simply be waiting to cross or they could be a potential suicide. By judging their demeanour and attitude from their behaviour and facial expressions we can assess their state and calculate the risk. These are all tasks that computers are incredibly poor at doing.
O L Leigh
I really cant see that being a problem. Any such driverless train will have an on board camera. This will use image recognition to assess what the obstruction is and decide on course of action. Image recognition is getting scarily clever. Also the picture will be sent back to control who would be alerted to an incident and can view and take action.
--- old post above --- --- new post below ---
It is widely thought that increased automation means less work for humans. Despite how much technology has moved on in the last 30 years, and how much automation has increased, it hasn't caused a corresponding increase in unemployment. The people previously doing tedious jobs that have been replaced by automation are now doing other things. So two jobs are now being done instead of one, which means increased productivity, and therefore growth in the economy.
Rich business owners do not want mass unemployment as they want people to be able to afford their products.
The problem is that low skill manual jobs are increasingly replaced with high skill technical jobs repairing and maintaining the robot.
This would be fine but for the fact that there are many people who only have the aptitude for low skill manual jobs and they are being thrown on the scrapheap ekeing a miserable existence on benefits
