Controlling the throttle and stick is pretty straight forward with bolt-on actuators. However, the force fed back has to be measured too. Commercial force feedb

Controlling the throttle and stick is pretty straight forward with bolt-on actuators. However, the force fed back has to be measured too. Commercial force feedback control systems for simulators cost ~USD20k per axis for a low backlash electric system with much less max force than a pilot can exert. Mounting this in on a robot arm with enough force to e.g. do a 10G turn sounds very pricey. There are many other controls in a cockpit that are not so straight forward to control. These frequently involve fine motor skills (like moving a trackball mouse to move a cursor). In larger commercial aircraft there are many many switches which are a full arms length reach away, or require standing (pre takeoff checks). A robot with the capability to do this is starting to sound pretty expensive vs a bunch of strap-on solenoids and actuators. Of course in an Airbus you could just plug in to the system bus. So really what we are talking about is having a non-human and non-manufacturer supervisory autopilot program. That isn't going to happen for passenger carrying aircraft until we have a much better way of writing programs to handle unexpected situations.