Obstruction detection for automated road vehicles doesn't work for trains because cameras and lidar can't see round curves to the limit of a train's braking distance. More fundamentally, the main line railway as it stands is nowhere near reliable enough for unattended operation and coping with breakdowns and emergencies remotely demands high capacity always-on track to train communications with built-in backups and 100% coverage. (GSM-R is a low capacity channel designed for voice, with less than 100% coverage and no backup). If the train fails or there is a track or OLE fault which immobilises it, you need to know exactly where it is (not just which signal section it is in) and that information needs to be available to controller and train crew and (if needed) emergency services. That's quite apart from monitoring the passengers, particularly as they board and alight and deciding when to close the doors.
Converting to the kind of automated railway you are thinking about demands a complete rethink of the operational concept, new rulebook, new safety principles, redesigning the signalling and control system to eliminate track circuits, new more reliable trains designed for ATO, self-monitoring switches and crossings, high quality radio comms - the list goes on and would cost £squillions. That's not to say it won't ever happen, but it's going to take a long, long time.
I was thinking that obstacle/obstruction detection would only be used for short range (183m / 200 yards) anyway. As on curved track, a human train driver is only using line of sight and can’t see round bends.
It may be possible for a longer range system to be used in future, once we have managed to work out how to remove background ‘clutter’ from being mistaken for obstacles or obstructions.
An obstacle/obstruction detection system is only ever going to be a last ditch effort if the train is travelling at high speed anyway.
I agree that the reliability of the whole railway system would need to be improved. Unfortunately I don’t see that happening anytime soon.
The responsibility for the passengers and the doors could return to being/continue to be the responsibility of the guard/conductor/train manager.
I’m not sure the track circuits, the points etc. are relevant here. It depends on which signalling system is in use. Track circuits or axle counters are likely to be retained as the final safety critical method of ensuring the track circuit block is clear. At least until more advanced systems become practical. In new installations point operating equipment is already monitored (*) in addition to the electrical detection. And in some places, said monitoring equipment is being or has been retrofitted.
(*) although how good it is, is another matter.
The on board computers on the train can pick up positional information from a GPS, radio data and track mounted Eurobalises. Plus use the train wheels to detect distance travelled and speed. Combined, this should give the on board computers a reasonably accurate current position.
By radio data, a better radio system would definitely be needed.
Yes, the procedures and rules would have to change. I personally don’t think it’s going to happen in the next ten to twenty years. It’s harder to predict beyond that timeframe.
But it is possible now (and by that, I mean the technology is available), even if it’s very unlikely to happen.
