MABEL Bipedal Robot is Fast Enough to Run You Down

reply to post by KayEs

Dude, this honestly scares the crap out of me! Does no one watch movies?! We're going to have a Terminator situation on our hands. There must be a way to shut them all down should that happen?

Anything that is made can be un-made with the correct application of force.

Joking aside - Movies are simply that - movies and drama. The reality of robots will be less ... stereotypical. For starters - we will be integrating digital electronics with our central nervous system long before we have the technology to make "robot armies" (in the context of Terminator or I, Robot). While autonomous drones do exist today - they are far more man-hour intensive than your average soldier; it's simply more practical to stick the gun in a man's hands and tell him to 'go take that hill.' It's much more difficult to get the robot to understand what you want it to do... and service its own weapon.

The main problem is power versus mobility. A lot of power is needed to run the computer systems necessary to run automated routines. In a platform like an aircraft or heavy armor (tank) - you have plenty of power coming from the engine (so long as there is fuel available), and some very powerful computing systems can be employed on those platforms with few consequences.

However... when you start to try and shrink that platform to something much smaller - to be more similar to a human, you find that the resources available to the design are in fierce competition for power. Mobility requires a lot of power with little room to store it. Computing requires a lot of power... and, again, little room to store it (and adding more weight requires more powerful mobility systems that consume more power).

That is why this design is tethered. Its power is all external (and it is difficult to tell if lateral balance is being maintained by the robot or the arm - it appears as though the setup is aimed at merely testing the leg design rather than attempting to develop a complete bipedal implementation). Attempting to run it off of on-board power would probably only allow for a few minutes of operation.

It would be a terrifying robot army for several minutes... then the centerpiece of dinner-table humor for the rest of human history.

reply to post by Aim64C


Very nice analysis! Solar power? you made many good points, I feel as if a solution COULD be found, which one, I am unsure of that.

reply to post by EvanJP


Solar power is really not practical for mobile applications. About the best you are going to get is a moving sidewalk with fixed solar arrays powering it. Even if you had 100% capture and conversion of the energy radiated from the sun, you would be hard-pressed to power anything we equate to a motor-vehicle, today (city traffic is too stop-and-go while highway driving is too power intensive). A military/defense application is out of the question.

The most promise comes from direct fuel-to-electric catalysis (much like a hydrogen fuel cell - only working directly with fuels like methanol or other hydrocarbons). That significantly reduces the weight of your primary power plant (the engine) and conversion (alternator) into a single solution. A similar application of cold fusion would also be viable - but we don't even have prototype/research devices to compare.

Then the only problem is in getting computing power compact enough to support anything we would equate to being intelligent. Even then... one must be careful to differentiate between 'simulated' and 'real' intelligence. It is a line that is not easy to draw... but, for example, let's say you have something like Watson... it's not an intelligent machine, it merely designed to give you the impression it is intelligent (it can service queries and match context to give you that impression, and even 'train' itself to be more accurate with future interactions).

Perhaps the best way to put it is the difference between intelligence and sentient intelligence. A sentient intelligence can be said to have a personality and will of its own. A cat... while not in possession of the same depth of intelligence of a human... is considerably different from Watson. I am sure someone will come along and chastise me for, essentially, declaring a cat sentient... a cat does appear to be self-aware. It is capable of understanding that there is an environment to be interacted with in different ways. Many even have a personality, particularly when interacting with other cats. Watson, however, is not aware of its environment in any context. It simply reacts objectively.

By time we do have computers and machines capable of replicating human intelligence, we will have already advanced considerably in cybernetics. I expect that some of the most popular cybernetic enhancements will start as interface devices and computational aids (a calculator that you 'train' to interact with and never have to search for, again). This will expand into 'notepad' type programs - and I expect, after that, 'rigid' cybernetic programs will dissolve, completely.

The reason for this is simply the way the brain works. Our neurology is very unique to each individual. The regions of the brain may activate in a very similar sequence when viewed under an fMRI when you and I are shown the same image... but the exact way in which that image is processed and how it triggers relevant memories is going to differ - perhaps greatly.

It will be exceedingly difficult to create devices and programs with a development process currently used for digital systems running on identical hardware. While we would 'train' to interact with cybernetic devices (even the simple calculator would require extensive training to use proficiently) - the way we currently develop computer code would require us to train for each -program- .

At such a point after being established, cybernetic devices will lose their rigid programming structure and focus more on developing devices that the user inherently 'expands' into. Rather than training to use a new program, you simply build one based off of the axioms you've trained to utilize. In a sense - your consciousness would simply grow to include the functions and operations of the cybernetic devices, much in the same way that people will often learn to think of or in reference to their arm, leg, or other physical peripheral when learning new topics.

So, by time we have sentient machines, we will have a much better understanding of how our own consciousness and sentience functions so as to better relate to our creations. Further, I am not sure the "robot versus human" issue would be as prevalent in such a world as it is, today... the distinction is great in today's world... but in another fifty to a hundred years - the distinction between a biologically-originated sentience and a technologically-oriented sentience may merely be academic (and with an increasing number of internet-based interactions, ideological differences and labels are more likely to encompass the greater volume of discrimination issues than physical).

reply to post by Aim64C


Thank you for the information!

I never assumed an all out "robot vs. human" scenario.. I wouldn't expect such advancements in A.I. so soon. I can see the remote control concept working very well, as coupled with humans perhaps servicing these machines on the battlefield, and keeping them energized? Trust me, I an NOT a fear mongering ass, I am not a doomsdayer or anything, I am glad that this has become a good conversation, and not ranting and raving about the robot matrix being built as we speak..

Glad its not just me who thought this thing was pretty creepy.. 100% the sound and gait did for me

Here is the big dog

reply to post by Freedom_is_Slavery


Haha, Only in this time of the obesity epidemic can a robot that runs at 10km/h run down a Human.

Originally posted by LightAssassin
reply to post by Freedom_is_Slavery

 

Haha, Only in this time of the obesity epidemic can a robot that runs at 10km/h run down a Human.

To be fair - it is theoretically possible for a robot to continue that pace nearly indefinitely, without rest, for very long periods of time. I'm not certain on the maintenance cycle typical of these classes of robots and what it looks like, or their manpower demand (probably 10 man-hours per hour of operation for these prototypes or so) - but I would imagine it will improve over the years.

The average human walking pace is, if I remember correctly, about 3 miles per hour... or roughly 5-6 kilometers per hour.

I can jog 'indefinitely' at about twice my walking pace and hold a solid 8-10 mile/hour run for twenty minutes or so (something I am working to improve - I should be able to hold that pace nearly indefinitely, but my core strength doesn't have the endurance to support both running and breathing for that amount of time just yet). In shorter distances, I get my kicks racing against cars in Wal-Mart parking lot that are going at least 15 mph.

So, yeah, this robot would have a bit of difficulty running me down (I only have to run faster than the slowest guy, right?). But it would become an issue if it was chasing me for, say, an hour. Of course, if it is still chasing me after two minutes, I need to be chastised for being too dumb to figure out how to get rid of it.

Originally posted by Freedom_is_Slavery
"For reference, MABEL's top speed of 3 meters per second probably isn't enough to catch a tolerably in-shape human, as Olympic sprinters can run at up to 10 meters per second over short distances. But the thing about robots is that they're determined, so in the end, it's a good thing that MABEL is tethered to that pole. And that it doesn't have any arms to grab you with. Or any vision sensors, either. So even if you can't run, at least you can hide."

War of the future is going to be #ed up

Only if they fight the war on flat level ground, in small circles, right next to power outlets. I vote for standing out of range and looking amused.