Once it's in place it's paid for you need to maintain it like building new lines or motorways. Cost shouldn't just be how much is it going to cost the railway, just like HS2 it's how it benefits the economy as whole. How many business outside the railways get affected by cancellations from railway staff absence thus affecting the economy. And look at the savings made you cancel the need for ''PNB's, they will work every Sunday ? How many DLR services are cancelled due to staff shortage compared to services that need a driver on each train?
SNCF are looking at trialling ''driverless trains'' in 2023. The higher the speed the more roles the computer and technology will take off the human inevitably to it being rather redundant.
Light rail and commuter services can be run closer together if run by technology they can react faster and run all day with a commands under the control of one person to a number of trains.
I think you are massively underestimating the complexity and cost of installing and maintaining automated systems. As well as the initial conversion cost, there will always be a significant amount of snagging, and then there’s the cost any time you want or need to make a modification, for all of which you’re essentially in the complete hands of software developers.
Look at TFL, who have a massive headache coming up when it comes to the Piccadilly Line - having boxed themselves into a corner with their ATC procurement for the current batch of lines, the favourite supplier knows they can essentially name their price, hence why Piccadilly Line resignalling is now looking increasing unaffordable for the foreseeable future. They’ve already had to descoped elements of the Sun Surface lines as a cost-saving measure where the benefits simply don’t justify the cost.
As for DLR, they run on grandfather rights, having been built as a small 1980s system, and massively expanded since. To build such a system from scratch, a platform edge door or obstacle detection system would almost certainly be necessary. Meanwhile, they certainly do get cancellations arising from crew unavailability, and they still have one of the biggest downsides of a manned system - the need for duty schedules.
The two issues with automation are (1) do the benefits outweigh the costs? and (2) is the technology ready yet? Outside of metro applications, the answer to both these questions is still quite clearly no, and the pace of progress doesn’t seem to suggest we’re likely to be anywhere near this changing in the near future.
.
