Originally Posted by
lucille
Between ILS airport pairs, the technology has long been here to fully automate. Now consider the immense complications of the Tesla autopilots being able to manoeuvre in a totally random environment. In contrast, the aircraft autopilot in a highly controlled environment is simplistic. Check out the Piper M600 Emergency auto land by Garmin. I’m not even sure it needs an ILS to guide it to touchdown. All this in a tin pot GA single. Imagine what Airbus and Boeing could do if they put their minds to it.
Everything an airline pilot does is either rule driven or can be made rule driven. And thus ripe for automation.
Pilot salaries are merely an annoyance to management. Their biggest constraint is rostering and complying with those pesky flight and duty limitations. And this is why automation is inevitable.
On the positive side bush flying in PNG and offshore Heli ops will be the last to be pilotless.
I don't really agree with that. Even the latest Tesla autopilot systems occasionally kills everyone on board when it entirely fails to understand its environment.
For vanilla normal ops, no doubt flight could be fully automated, but how often do you get a perfectly standard ILS to ILS day? Hows automation going to deal with the countless variances that in my view, require a decision based on inputs that aren't easily translatable to a machine? Strange smell on the flight deck? Dangerous weather return vs just heavy precip? Alarming noise during the t/o roll? On the 787 there are dozens of unannunciated checklists, ones that are only executed after a human decision because the aircraft has no idea what is wrong. or even that anything is wrong at all.
I think there is far, far more than needs to be developed for automated flight to occur, and I suspect its a long way off at this stage. My examples are just a drop in the bucket of the challenges that automated flight faces. And given the development time in aviation.. well.