Japan aims for walking robot on the moon by 2020 Options

[lyford]

No word if it will be a Transformer

Japan aims for walking robot on the moon by 2020

TOKYO (AP) — Japan hopes to have a two-legged robot walk on the moon by around 2020, with a joint mission involving astronauts and robots to follow, according to a plan laid out Friday by a government group.

Also, if it is a humanoid machine with artificial intelligence, at what point do we stop saying it is still unmanned space flight?

[Geert]

I have often wondered about this.

If Spirit would have had legs instead of wheels, I guess she would have far less problems in this type of terrain, and there is a lot of very interesting terrain on Mars which is simply in-accessible to wheeled rovers.

Still I would prefer 4 or even 6 legs (far more stable then 2, and you can always make them in such a way that they can double as robot-arms).

Problem with legs is simply the complexity (=weight/costs) of the whole thing and durability, you have to use pneumatics or hydraulics to make this work and that will give a big penalty in weight plus all those moving parts will need to keep working for a long time in a very harsh environment (not impossible though, there are lots and lots of complicated hydraulics on earth which are running for years and years in very harsh environments). Energy-consumption will be substantial as well with legs, I guess, a lot of problems still to solve but in the end imho, just looking a mobility, a 'walking' robot will be better then a rover.

[dvandorn]

I have a bad feeling about this -- look at how much trouble the Japanese are currently having making Asimo walk up and down stairs.

Bipedal locomotion is not intuitive or simple for machines. Add in the entirely different ways in which mass responds to force applied in a lowered G-field, and I'd really hate to see their walking robot take two steps, fall over, and not be able to stand back up... *sigh*...

-the other Doug

[nprev]

Definitely don't think that bipedal locomotion is the way to go for planetary surface exploration robots, anyhow; it works for us, but we've got well-developed hazard avoidance mechanisms & means of geting ourselves out of falls.

Hexapodial: there's the ticket, IMHO. You get the versatility of walking, plus six-point stability (four doesn't seem like enough for extreme angles & winds without sophisticated balancing features, and eight is overkill in terms of unnecessary control complexity).

[lyford]

Hexapodial: there's the ticket, IMHO.

Now where have I heard that before?

[Geert]

Now where have I heard that before?

Very interesting design, but I doubt a bit why you would need the wheels on the end of the legs. The thing is designed to cross terrain where you can't use a rover, that's the whole point, putting wheels on the legs only makes it more complicated with even more items which can (and will) fail.

It won't need to go fast, if it can set an 'easy walking pace' it already will outrun every present rover, speed is determined by how fast your computer hazard avoidance system can assess the situation and even with the best computer techniques it will probably take quite some time before there is any 'need for speed'.

[Phil Stooke]

There are a lot more actuators in a leg than in a wheel. This is a disaster waiting to happen.

Phil

[imipak]

Hmm, the actuators on Mars today seem to be performing pretty well. Even the grouchiest, grumpiest, most cynical "meh! It'lll never work" prophets of doom (...that'd be me) like to have little dreams of future possibilities now and again. Presumably it won't be attempted until the risks of disaster are deemed appropriate.

[Paolo]

remembers me of UFO Grendizer...

[nprev]

Phil's right, though; moving parts should always be minimized, and that's the very definition of actuators. Still, as the state of the art advances, there may well come a time when a 'walking rover' might be the best solution for reaching high-value difficult-access places of interest on Mars.

Remember the rock gardens of Chryse (V1 & Pathfinder) & Utopia (V2)? The MERs would have had an exceedingly hard time getting anywhere at all had they had the misfortune to land in similar places.

One underappreciated pioneering aspect of the MERs, BTW, is the amount of way-beyond-the-design-envelope testing that they're putting their own actuator suites through almost every day. Bet that there are some vastly improved designs to come in terms of reliability from their lessons.

[Geert]

Bet that there are some vastly improved designs to come in terms of reliability from their lessons.

I think there are already actuators in industrial applications which work for years and years in environments which are far more extreme then the surface of Mars, however these are mostly far too heavy and far to power-hungry to use on a Mars spacecraft. It has been stated many times already but each spacecraft is a compromise in how far you can get in making it perfect and still remain (more or less) within budget.

I have no doubt it is already possible to design and build a 6-legged robot which is rugged enough to walk for many years on Mars, however you will need a Saturn 5 to launch it to Mars and I don't want to think about the costs...

Still, if you look to the HiRISE images, in the near future we will need something which can handle much rougher terrain then MER/MSL if we like to explore some really interesting places..

[MahFL]

I just don't see how 2 legs can be safe. If you think about it humans have 4 appendges, not 2 legs. We of course used to walk on 4 legs.......we only evolved to bi-pedal because that is what was needed to survive and evolve due to our enviroment ( open plains I beleive ).
So bottom line is, min of 4 legs makes sence.

[remcook]

This wasn't released in Japan on April 1 by any chance, was it?

[cbcnasa]

I agree with an earlier statement, ASIMO looks great in the commercials but has problems on common surfaces. I would like to see it but I think the cost in getting it there will be so hight.

[PDP8E]

Mankind has a million years of experience with bipedalism. We also have made 60 billion bipedal prototypes (the approx number of all humans that ever lived).

Wheeled vehicles have been around thousands of years. The auto industry has cranked out almost a billion units in the last century.

ASIMO has been around a decade and half and is probably manifested in less than 500 units (mostly prototypes).
I have never seen a video of an ASIMO-type bot recovering from a fall. (although I did see one riding a bike!)

Although I am personally skeptical about an ASIMO-like bot on the moon, we also know that a lot of things can happen in a decade (2020) so, Japan, good luck with that bipedal moon-walking-bot thing, we will be watching.