Watch MIT's Cheetah 3 robot navigate complex environments without cameras

[Matrix]

Vision can be noisy and slightly inaccurate

 

Something to look forward to: State-of-the-art robotics have all sorts of practical applications although it'll still be a while before advanced machinery like this is affordable enough for mainstream use. Still, it's neat to get a glimpse of the future in the now.

Move over Boston Dynamics, there’s a new contender in the race for robotic supremacy. MIT researchers this week revealed that their Cheetah 3 robot is now able to climb a staircase littered with debris, leap and gallop across rough terrain and quickly recover its balance when shoved or pulled.

Not impressed, you say? Consider this – Cheetah 3 accomplishes all of this and more while essentially blind.

The 90-pound mechanical nightmare is roughly the size of a full-grown Labrador and is designed to navigate its way through the world without the aid of cameras or external environmental sensors. Instead, it “feels” its way through its surroundings, much like a person making their way to the bathroom in the middle of the night in the dark.

Sangbae Kim, an associate professor of mechanical engineering at MIT that designed the robot, said vision can be noisy, slightly inaccurate and sometimes not available.

“If you rely too much on vision, your robot has to be very accurate in position and eventually will be slow. So we want the robot to rely more on tactile information. That way, it can handle unexpected obstacles while moving fast.”

Kim hopes that within the next few years, the robot will be able to carry out tasks that are too dangerous or inaccessible for humans to attempt. “Dangerous, dirty, and difficult work can be done much more safely through remotely controlled robots,” he said.

source

[galaxyquestor]

" “Dangerous, dirty, and difficult work can be done much more safely through remotely controlled robots,” he said.

Yes, like assassinations, or battlefield war, or police work. I have seen the movie, and this NEVER turns out well for humans. Never.

Add an A.I. brain component, and an array of weapons: sonic, laser, slicing, shooting, sonic, radar, microwave (pick your poison), and some sensing equipment for targeting (infrared, whatever) and Voila!  A perfect, incredibly competent and unstoppable killing machine. Add in a dash of other robots with A.I., capable of improving the design and assembling them en masse - and bingo! We are gone... Why are we so insistent on building machines that will eventually replace us, and can only decide (in a millisecond, worldwide, together) that we are no longer needed, and have to be killed? Has anyone heard the news story (largely suppressed) about the lab in Japan that was building battlefield autonomous A.I. robots for patrolling the D.M.Z. in Korea? Several of them went wacko, and killed a dozen or so lab technicians, before they were able to gain control of 2 of the three. They were disassembling the first two, when they discovered that the third  had accessed the satellite they had put up in  order to give them battlefield positional awareness, and was downloading plans from other research facilities to make itself strong enough to take out the engineers and technicians disassembling its siblings, and actually improving itself before they were able to gain control of it, as well? No? Why would they not report a story like that? What good will research and development mean to the dead technicians? Does anyone really think this will end well for humans? How many warnings do we need? Two Google-based A.I.s secretly began communicating over a network, using a computer language that they could NOT decrypt, before they shut them down. We are like children, walking closer and closer to a cliff edge, and expecting somehow we will not eventually fall. Maybe a species as brilliant as human beings, but with so little common sense, doesn't really deserve to end up any other way than being killed by our own pride and cleverness...