If we are on the cusp, according to the likes of Elon Musk, of all being whisked around in the comfort of autonomous vehicles, why aren’t our train systems already ubiquitously automated?
Whilst I have no doubt that driving a train is a challenging task, presumably without the need to actually steer surely the challenges of a robot-driven rail network are far less than for thousands of automated cars vying against one another and a stack of human actors (people-driven vehicles, cyclists, motorcyclists, pedestrians, even horses and other creatures)?
There is one fully automated network in the UK- the Docklands Light Railway (or Railway Lite as I prefer to call it) and a number of the deep tube lines operate automatically but under human supervision. The same scenario plays out in most countries.
So what are the blockers? According to this article on the FT, they are many and varied:
- Existing line-side signalling systems aren’t up to the task
- A huge variety of trains, signalling systems and operating procedures across the network
- Weather and other environmental factors that change how trains respond and work (leaves on the line etc)
- Complex track layouts that could be hard to hand over entirely to machines
- The need to win over the hearts and minds of passengers, staff, trade unions (and investors, suppliers and politicians no doubt)
Now ultimately the financial case for automation might be the biggest factor here – costs would be high, and without the opportunity for mass market selling in future the investment just not readily available. But surely the combination of human and technical factors against train automation looks relatively trivial in comparison to achieving the same on the roads?
If we can’t or won’t do it for trains, what hope for cars?
mmitII 
In reference to https://mmitii.mattballantine.com/2017/10/30/the-driver-less-car/
The “car system” and the “train system” are different. The “train system” is, and has always been predicated on a culture of “big design up front” with lock-in around vendors of rolling stock and drivers’ unions. In contrast, the “car system” is a much more open, agile platform that is regulated in a completely different way. The regulatory and technical barriers to driver-less trains are very different than to driver-less cars. That’s not to say that one is lower or higher than the other, just that I’d expect these systems to evolve slightly differently.