Pat Flannery writes:

NASA's "Project M" tele-operated Moon robot video:
http://nasawatch.com/archives/2010/0...jscs-proj.html
NASA says this can be done in 1,000 days from the word "go".
Is there any good reason it needs to be this anthropomorphic?
I can see the head, torso, and arms...but having it actually walk around on
legs rather than using wheels? The legs would have to be automated somehow
because of the time lag in communicating with it to prevent it from falling
over.

This is interesting. I'm not sure about the 1000 days, but I'm not privy to
what JSC has been up to.

Anthropomorphic makes sense when you think about it from the controller's
perspective. I sort of know how my autonomous and non-autonomous nervous
systems respond to two legged and two armed systems. I'm not sure as a
human being how I'd best teleoperate an eight-legged system or even a
six wheeled system.

Think of it from the tele-operators perspective.

As far as the comm time lag goes, I'd mentioned a good way to deal with that
many many posts ago, using active resistance feedback to provide a way to slow
our own ground based motions to match those of the device on the moon.

Another item to consider. Sending up lab equipment so that our robot can do
experiments on the moon. It might be easier/cheaper/simpler to modify existing
Earth based human operated lab equipment to deal with an anthropomorphic robot
on the moon rather than start from scratch with some kind of eight legged, four
wheeled or even snake like robot.

Also think of it as a way to get good at designing equipment that eventually
could be dual-purposed and used by humans on the moon as well. This is the
back-door way back to the moon that doesn't hog up all of NASA's budget trying
to do it. I like it.

As for active balancing, put the gyroscopic feedback into a tilt table the human
has to ride back on Earth. Or better yet a 3-axis moving platform the operator
'rides'. The robot falls down, the operator pitches forward face down in the
three axis harness.

Dave