Dunno why they don't just go ahead and bite the bullet and go to 100% in-cab signaling. The technology has worked for decades and has been deployed on many lines, including a few in the UK.
Cab signaling requires a different set of standards in order to provide the necessary level of safety.
Cab signals are (usually) transmitted through the rail and picked up by a special receiver on the leading axle of the locomotive. Each locomotive has to be equipped and the rails themselves have to be checked and blocked. This also means the rails have a higher maintenance requirement as opposed to lineside signals which can be more easily maintained.
However, some countries (Brazil predominantly) have embraced a GPS based cab signal system where lineside signals are almost totally eliminated. However, this was done with federal subsidies to cover the expense iirc.
Another example is the US Northeast Corridor where cab signals are used along with lineside signals to provide a level of redundancy at speed.
Unfortunately there’s no easy solution, but I lean towards the Brazilian GPS based systems from Wabtec.
Realistically, in the UK, it's going to be ETCS with balises: all new build stock at this point either has ETCS in-cab signalling fitted, or is designed for its fitment.
That said, the rollout of the infrastructure side of ETCS is somewhat delayed, but it is happening in places (Thameslink is ETCS Level 2 with conventional signal overlay, albeit with shorter blocks in ETCS; Crossrail is ETCS Level 2 with no overlay).
As a random piece of trivia, in cab signaling was pioneered in the 1920's and implemented commercially by Pennsylvania Railroad the 1930's while upgrading their main line between New York City and Washington D.C. You can see in this photo that the engineer had one indicator that would light up for each signal aspect: https://i.pinimg.com/originals/31/85/83/3185833fba11b79498ff...
The locomotives, not all the rolling stock. I'm not saying it wouldn't be hard/expensive, but it seems better to spend money on things that will be useful for decades rather than stopgaps that just kick the can a little further up the road.
If you remove the block signaling system and use in-cab signaling, you still need to know where the back of the train is. Otherwise a train could run into the back of another in the fog or heavy rainstorm. Currently, railroads use FREDs (flashing rear end device) on the last car of the train. Some FREDs are more intelligent than others but would benefit from being able to send out their location via GPS.
Longest trains are under five miles so you could use that or whatever the limit for that line is as a safe baseline. Unless you're worried about trains separating.
That's not an actual problem. You simply have 2 trains each broadcasts it's location. It's unexpected separations that might be a problem, but those are currently very rare and dealt with seperatly than static signals.
It's like you need to design a nuclear reactor to deal with an aircraft hitting it. You don't need to automate the response becase that's a rare situation.
Well, I read about one instance where it happened, and the brakes didn't apply. (Not sure how that could happen other than a mess-up in building the train. You'd just about have to have both angle cocks closed.) Both fragments continued, for miles.
Can't cite a source (I think I read about it on the Altamont Press discussion board). And that's only one instance. And it could have been reported incorrectly. But it appears to be possible for this to happen.
Most British passenger trains are multiple units (and most British trains are passenger trains), so there's effectively little distinction between locomotive and rolling stock.
But when the people who hold the pursestrings are looking no further ahead than their next election, "useful for decades" means nothing at all to them.
