Some more thoughts having followed the posts which followed on from mine.
Firstly - the ultimate goal of technology is the removal of humans. Not true - technology, like everything else evolves. It is Darwinian. There is no grand plan, what happens is the result of many people, not necessarily in the same company, the same country or even the same continent, trying out ideas to improve something. Then these ideas come together and somebody sees a market opportunity. Sometimes these things work, sometimes they don’t. There were no ‘grand plans’ to develop the motor car, the mobile phone or the internet - someone came up with an idea for an internal combustion engine which was the ideal weight, power and size for a horseless carriage and, as it happens, aeroplanes. Nobody started out with the idea that they wanted to make farriers, grooms, tackle makers and stable boys redundant by inventing the car. Or that they foresaw out of town shopping centres. Closer to home on the railways - there was no ‘grand plan’ to develop railways, but various technologies - metal working, ‘strong steam’ (as distinct from the atmospheric engine) and a market need to move goods more effectively came together. When it was showed that the idea worked then railways exploded across the landscape - but even a few years after the Battle of Waterloo there was no concept of such a thing happening.
Secondly - the type of information support/device I wrote about in my post was not intended to control the train, but supply support to the driver. However there is no reason whatsoever why it could not directly control the train - and it could control it in realtime using feeds from the railway control centres and traffic management systems to meet whatever service requirement is most important at the time - minimise delays for all trains, ensure connections are made, minimise delays for the train in question. All this is possible - and it is not expensive, practically all the data for such decision making are already available.
I assume that, even at some time in the future, passengers will not be too happy with the concept of driverless trains so a driver will be present. At this point the issue then becomes one of the best way to keep the driver alert and be in a position to take over - if he/she only supervises the electronics that becomes boring or does he/she drive the train but the computer takes over if it anticipates a problem. This is the same problem that faces the designers of aircraft flight control systems and monitoring equipment for industrial processes such as chemical works and oil refineries.
Thirdly - the concept is risky or unreliable. Just saying that something is risky, especially with no explanations as to why, does not make it so and is no argument against the change. Of course everything can, and probably will go wrong - and this is as true of humans as well as of technology. However it is possible to design systems which meet any required level of accuracy and reliability; for the system to be safer than humans it only has to be a bit better. Somewhere I read that there is one SPAD for every 50,000 (or some such number) signals approached showing a red aspect. For railway safety these are pretty good odds as most SPADs are overshoots of only a few feet and are well inside the overlaps. It is only necessary to design a system which SPADs once every 100,000 approaches to a signal at danger and the railway is even safer. But the real benefit for the railway is not in safety, but in reducing the delays which inevitably follow a SPAD: all the reports and paperwork and testing for drugs and so on. In this case the computer also reduces the chances of a black mark on the driver’s copybook.
The actual technical problem is not that something will go wrong but how does the system degrade when it does? Aircraft flight control systems are frequently triplicated, if one channel gives a spurious signal a a voting system choses the two signals that agree and ignores the third one. (This is terribly oversimplified but gives a flavour of the approach). This is necessary because if something goes wrong in an aircraft you can’t, in the words of the old song, ‘get out and get under’. If there are discrepancies anywhere in the driver support system under discussion it is always possible to stop the train and work out what to do next.
Fourthly - such a driver support system needs ETCS. Not true - as long as there is a digital model of the route a suitable display can be produced showing the features along the way even if the lie of the turnouts can’t be displayed. If it was decided that the railways need a location system independent of GPS (as its accuracy can be degraded in times of crisis or war) then all it needs is that all routes to be equipped with ETCS balises. These are essentially Radio Frequency ID chips in a plastic box - they are passive and need no power supply and are cheap. They can be added to any route independent of the type of signalling as they only act as location beacons for the train.
Fifthly - for any technology aimed at a particular user group the needs of that group have to be taken into account. The designers are specialists in their field but they are not specialists in train driving. So it is easy for them to get it wrong - either it won’t be effective as it solves a problem which doesn’t exist or it will be difficult to use and won’t be accepted at all. But bear in mind what the driver needs now may not be the same as what the driver needs in 15 years time. The system also has to evolve - Darwin again!
All change is unsettling, especially if one’s job or livelihood is perceived as being at risk. As in all things there are winners and losers - and it is often not possible in advance to predict who will land up where. But change will come - it always has and it always will. The trick is to work out how to use it to one’s own advantage - and history shows that simple opposition always fails. My feeling is that if used correctly it will enable some more trains to be run on any given set of rails. This will be beneficial for the railways as it will enable, in some areas, more traffic to be carried before very expensive infrastructure enhancements become necessary. More trains translates to more drivers.
