From that era, and from what you know so far (that they are apparently
pretty much all germanium parts), it shouldn't be too difficult to
deduce how it worked. It's unlikely any of the active parts are
particularly esoteric. Usually you can figure out the polarity of the
parts from the schematic, and assuming about 0.3V for forward junction
bias and current gains typical of the power and signal transistors then
in use should get you close to what you need. I suppose nothing in an
autopilot needs to be very fast, so it's very unlikely any are RF
transistors. It's possible that the circuit requires close tolerance
on some feature (like gain), but that's poor circuit design and I'd
hope for better of them. Once you have a circuit diagram, maybe you
could try "spicing" it, if you're not comfortable with a simple paper
analysis.

Good luck with it.

Cheers,
Tom