I'd do some reading up on the project before making posts like tht because scenarios such as those you describe have been planned and catered for. The 700s and core platforms will have level disabled access. ATO will not have any control over closing doors-all ATO does is allow the train to practically run bumper to bumper with the train in front.
Also 4-tracking St Pancras won't make a difference as it still goes 2 track further into the core. If all goes wrong then St Pancras high level and kings cross are both available to turn trains around and thin the service out.
I'm not saying it will be perfect but all of your concerns have been being addressed over the past few years and are still being worked on. Things like disabled people boarding and alighting are not issues to worry about as they have been catered for.
I do wonder at times that the benefits of ATO can be a little overstated. The main benefit is every train should be driven consistently, designing out the problem of different drivers driving in different ways.
However, using the example of one London Underground line which has recently gone ATO, when waiting to enter a platform, the new system often stops the train further back from where the outer home signal used to be, and when the train ahead starts to move out of the platform the system waits longer than under colour-light signalling before giving a limit of movement authority. Often the train is still standing still at the moment when under signals all three home signals would have come off.
The limiting factor is nearly always passenger behaviour. 24tph is certainly achievable. The Victoria Line, for example, reliably achieves 34 tph during the peaks. It will be interesting to see what happens at some of the Core stations during the peaks - the reality is there are going to be intervals to some destinations of 30 minutes. Some of these people are going to be anxious to stand in the particular spot where they know their doors are going to stop - they will be keen to secure a seat, not helped by the class 700s having fewer seats than comparable formations today. Great Northern commuters are already well versed in this at places like Hitchin and Stevenage. What effect this will have on train dispatch will remain to be seen, but if it causes a problem it's a behaviour that will be very hard to stop.
For me, no one has yet satisfactorily addressed the concern of multiple service groups intermingling with each other, and the effect this will have on performance. Some have said performance will not matter so much because frequencies will be "turn up and go", but this is not really acceptable as most services will be half-hourly, this is certainly infrequent enough that people will expect to turn up at the advertised time and the train arrive on time. Despite the infrastructure works carried out or underway, there will still be loads of places were small amounts of late running will introduce conflicts. The long double-track section between Hitchin and Cambridge will host three different stopping patterns with a difference in running time of up to 15 mins between the fastest and slowest services - as well as conflicting with other services in the Cambridge area. The slow lines between Welwyn and Potters Bar will also host different stopping patterns with the potential for slow services to delay faster ones. Meanwhile, we have two long single-track sections north of Littleport which will be carrying an all-day half-hourly service from 2017 - twice that of today.
No amount of modelling can predict what will happen in the real world, as we have seen at London Bridge recently. And no amount of modelling can resolve the problem of two trains arriving at a converging junction at the same time, or a fast train catching up a late-running stopping service. Only additional tracks can expand capacity in this way; apart from a bay platform at Stevenage to keep the Hertford services segregated, the Thameslink Programme appears to deliver nothing of the sort for the Great Northern network.