Richard Riley writes:

It is perfectly capable of flying an entire mission, from engine
start, to cruise flight, to touchdown and engine stop, autonomously.

That does not surprise me. I recall reading about proof-of-concept studies
along these lines when I was a kid, and I'm sure the field hasn't stood still
since then. The principle is pretty straightforward, although the devil is in
the details, as usual.

In the US, at least, we are required to have a set of mark 1 pilot
eyeballs on it at all times to maintain separation. And the pilot has
to have a class 2 medical and a commercial certificate.

So the pilot is a regulatory requirement ... not a technical one.

Visual separation is only needed because other aircraft in the area are not
autonomous and in communication with yours. But imagine, say, Class A
airspace in which all aircraft are similarly equipped. Nothing up there but
other aircraft of the same breed, and no obstacles or terrain. It would be
much easier to fully automate things there. Later, it could be automated at
lower altitudes and in different flight phases.

Unfortunately, this implies either severe restrictions on small GA aircraft,
or extremely expensive upgrades to those aircraft to make them compatible with
automated aircraft. The other option--making autonomous unmanned aircraft
capable of maintaining visual separation--is probably not feasible technically
and economically.

Already, RVSM requires autopilots. The trend will probably continue, with
pilots becoming more and more peripheral to the flying task.

There may come a time when the only real flying that pilots will be permitted
to do is private GA. All commercial aviation will be handled by robots, and
if human "pilots" are present at all it will only be as emergency attendants.

In some ways, this is easier for aviation than for road vehicles, although
it's easier still for trains, I think.

--
Transpose mxsmanic and gmail to reach me by e-mail.