Nurse Ratchet... Literally!

I was just at a conference where someone was describing tele-surgery along these lines. The trick for anything that requires real time control in situations like this is how to handle communication delays. When I was in grad school we did a lot of experimentation with master-slave manipulators to investigate the effects of time delays, and the results were sobering. For delays of more than a second or two (as I remember) people adopted a move-and-wait strategy where they made a small move then waited until the visual results of the move came back to them. Hardly the way to perform surgery (or to experience it!). Bummer to end up with your lower lip stitched to your left eyelid because someone was dowloading Everybody Loves Raymond reruns on a nearby network...

I think that it only works in cases where the person will die if nothing is done. The lag could be a killer, literally. I'm thinking that a satellite unlinked remote surgery unit would be much more likely to be used in remote locations throughout the world where an expert in the surgery observes and instructs a less experienced surgeon.
If I'm not mistaken the move and wait approach is what NASA has to do with the rovers. The signal go to Mars, the action to be taken and then for the signal to come back. It's not life and death, but it must take patience.

Yup - this is the rover strategy as well. A lot of work along these lines has been supported in the autonomous vehicels arena by the Office of Naval Research. Their interest dates back decades (and supported virtually all of the Man-Machine Systems Lab at MIT while I was there) because cameras and other sensors are bandwidth-limited underwater unless you use a tether. The whole supervisory control idea blossomed out of this need to tell moderately intelligent systems "go here, do that, don't get hurt, and let me know how it goes when you're done". We naturally pull back from the full game when it's a multi-hundred million dollar program but I think they're doing some of that on Rover.

It's interesting to see how smart our "moderately intelligent systems" really are now a days. I doubt that even 10 years ago people could dream of the complexity of the robots that we are sending out to explore our solar system.
Our rovers on Mars are well beyond their expected shelf life, but still going (fairly) strong. It seams that NASA has a knack of over engineering everything; to the benefit of science and man.