Hyperion Rover, 1 km On One Command

An anonymous reader writes "Carnegie Mellon's next generation robot just finished its Chilean expedition and achieved a new planetary exploration benchmark, including being the first autonomous rover to cover 1 km on a single command. The other milestones from the Atacama Desert, Chile--the driest place on the planet--centered on over-the-horizon stereo navigation, sun-tracking for efficient solar panel pointing, and fault recovery. CMU shows pictures of the robot, called Hyperion, in action. One of its prime objectives was to plot courses that avoid shade, by finding the position of virtually everything in the solar system."

Comments

Re:That's impressive

[in7ane]

That's true, from the photos it looks like it just rolled along a pretty much flat field.

The setup simply looks too flimsy (assembled out of two bikes?) to go over any significant obstacles.

I do not doubt that the tracking system, etc. are impressive - they've just got to hook them onto a decent base and send it across a more challenging landscape. On the other hand that will seem like a military application then :)

cool !

[jacquesm]

This is really cool... but other than that the general state of robotics is more and more hampered by legalities, such as liability insurance for the 'owner/operator' of any kind of robot.

It's funny how if you look back at the turn of the century there was no legal barrier to try out new technological stuff, but just try to imagine the then inventors of automobiles selling their first rickety inventions in todays unbelievably hostile legal climate. The whole technological and transportation revolution would simply not have happened

That's why we see robots for use on other planets, but we'll probably not see them on this one (unless of course we ship all the lawyers to some other planet first).

Re:cool !

[jacquesm]

no, my premise is that it is no longer possible for smaller companies to bring stuff to market because the first (even failed) lawsuit will probably put them out of business, and because insurance companies will not cover product related liability at a price that will make the product still affordable because of the absolutely ridiculous amounts that are routinely awarded in product liability cases.

Around the turn of the century people accepted that new technologies and their development incurred a certain amount of risk taking on the part of the public, nowadays we expect to be absolutely safe from the cradle to the grave.

Lawyers have heavily capitalized on this (especially in the US) with all kinds of bogus lawsuits about product 'failures' (you probably know the various examples as well as I do.)

This is keeping a whole pile of potentially interesting devices of the market or from being developed at all because the would be developers feel that actually selling their work would expose them to all kinds of harm (especially if they are somewhat successful).

The 'I'll sue you' attitude is becoming more and more widespread and is having an ever stifling effect on development.

The only way we are going to go forward is by making mistakes (accidents) and taking risk, not by taking a risk, getting into an accident and then to sue left right and center just to either make a buck or to soothe some inner child that feels wronged.

Large corporations see the new legal culture as a new form of tax, smaller corporations simply go under (or never even get off the ground).

It's evident that robotic technology will initially at least lead to all kinds of exposure to risk, especially if we let the devices loose in our urban jungles. But if we do not then we'll never learn what to fix, and the development will be slowed down to the point where you'll be hard put to mark any progress at all since a device will have to be absolutely bullet proofed before it can be sold.

Re:Sun Guided

[JJ22]

I think that would have to be referring to the ability to detect *locally* where the best place is to stop as sundown approaches (such as not going into a valley if the sun will go down before it can get to the other side, or stopping on a hilltop to catch the earliest rays in the morning).

Re:It has to What Now?

[frovingslosh]

Yup, the same statement struck me as rather absurd too. No only would the stuff you track be meaningless, but say, for example you were the little robot that could exploring Mars and you calculated that Earth was about to eclipse the sun (not that it could be much of an eclipse from Mars' view, but you must be tracking these things for some reason). What the heck can the rover do? It's not like it's going to move to avoid the eclipse!

And yes, I read the link, it says nothing about this. Perhaps submissions by anonymous deserve a little more editing.

Re:It has to What Now?

[GooberToo]

It could go into a power saving mode until the eclipse has passed. If power isn't being generated after the eclipse has passed, it then could begin trying to figure what the heck happened. It can be used for any number of intelligent diagnostic and power saving modes. Also, if it knows all valid times that its cells should be getting power, it can figure out if it navigated into shade or if it's one of the other known events. If it's a known event, it may be able to wait it out. If it's an unknown event, it can attempt to backtrack and report a possible problem.

Remember, just because it tracks various bodies, doesn't have to mean that it does so in real time.