> Not really much point to this but I thought it was interesting how the PSAs are effectively put in a position of responsibility for the actions of the autonomous vehicle but presumably are not compensated the same as an active(?) driver would be.
This is an interesting question. Most drivers on the underground aren’t active drivers at. I think every deep level tube line, except the Bakerloo, is autonomous, and the trains can’t be driven manually in normal operations because a human driver can’t react fast enough to continuous adjustments that are sent to train. Adjustments that are needed to safely operate the tube system at the frequency and close follow distance or achieves during peak.
Instead the only things a tube driver does is check the doors before closing them, instructing the train to depart, and hitting the emergency stop if something goes wrong. They probably do get paid more than PSA, but I know they get extra compensation due to amount of time they work underground.
I do want what level of liability a PSA is exposed to in that position. Presumably they would be in trouble if they were obviously not paying attention during the evening hours. But assuming they’re playing in their phone, I suspect they have little to no liability at all in the event of an incident. But I suspect that the emotional toll if “their” train hit someone would be high, regardless of their responsibilities.
> Maybe a lot of the desire to move to "autonomous" (or really just more automated if it requires a safety driver) is a workaround for union wages by reclassifying drivers as new made up things.
TfL actually have no desire to automate existing lines. The costs involved are astronomical, the saving minimal, and service improvements almost non-existent. For TfL there number one priority is improving capacity, and the biggest blocker to capacity on the underground isn’t drivers, it’s trains, signalling systems, and above all dwell time in stations.
The signalling systems place a hard cap on maximum trains per hour, and the existing systems haven’t performed as well as hoped (it should be mentioned that these systems were cutting edge when installed, TfL tech tends to be on the absolute bleeding edge of metro technology, but the problem is fundamentally harder to solve than original expected), and dwell time (the time spent in stations getting people on and off the trains) his hard to reduce much further, without re-engineering trains and stations to increase the number of doors and platform crowding.
Central government has been pushing the automation angle because it’s a “simple” solution to TfL finance problems. But as usual the problem is vastly more complex than that, and the “simple” solution is only good because it feel intuitively correct to public that doesn’t understand the actual problems TfL faces.
London reconnections has a number of fantastic articles covering all this:
https://www.londonreconnections.com/2014/driverless-trains-p...
https://www.londonreconnections.com/2020/tfl-the-impossible-...