Humanoid Robot Brains (Science Out Loud S1 Ep2) - Free Educational videos for Students in K-12

Humanoid Robot Brains (Science Out Loud S1 Ep2) - By MITK12Videos

Transcript


00:0-1	when you jump over a puddle , climb a tree
00:02	or play a game with basketball , you're doing things
00:04	that no robot can currently do your outperforming the work
00:08	of thousands of scientists , companies , and even governments
00:12	. Now , for you and me moving around really
00:15	easy , but for our robot to control its motion
00:18	, it's much harder if we can somehow copy our
00:21	motor control and robots and we could save lives by
00:24	replacing humans on hazardous jobs like firefighting , performing a
00:27	spacewalk or even cleaning up the meltdown at the Fukushima
00:30	nuclear reactor . If I want to grab this drill
00:35	, I just pick it up . There's no thinking
00:37	involved . But why is that ? It's because my
00:41	subconscious mind is recruiting billions of brain cells to do
00:44	all the complex calculations involved . For me When I
00:47	decided to pick up the drill , I first rely
00:49	on 30-50% of my brain to subconsciously process the signals
00:53	from my eyes now in the front part of my
00:55	brain . I consciously whip up a plan to reach
00:58	my arm out towards the drill , touch the drill
01:02	, grasp it and then bring the drill back to
01:04	my body . My brain subconsciously transforms these vague commands
01:08	. First two basic motor plans in my pre motor
01:10	cortex , then more precise muscular instructions in my motor
01:14	cortex . Finally , my cerebellum perfectly coordinates activation patterns
01:19	that get passed to the brain stem and sent out
01:21	to all of my muscles . Once I take my
01:24	first step , my brain is constantly comparing the plan
01:26	with the real time situation and adapting to what I'm
01:29	sensing . Even if I'm not fully aware of it
01:32	now , you can start to see why it's so
01:34	hard to get robots to do these complex tasks .
01:36	It's not as easy as just controlling a motor .
01:38	You have to plan and perfectly adjust all of the
01:41	motors together with robots that have wheels for trends ,
01:47	we don't have to worry about balance since we always
01:50	have a stable base . But if one of these
01:53	robots comes across stairs or a ladder , it's michigan
01:56	come to an early end . That's how we're working
01:58	on bipedal robots . The lots that have two legs
02:02	though , they're harder to control their more adaptable and
02:05	better suited to work in environments that were made for
02:07	humans . I'm here with a bipedal atlas robot made
02:11	by boston dynamics that the M . I . T
02:14	. Team is using to compete in the 2013 DARPA
02:16	robotics challenge . You can call him Helios . Now
02:20	Mitt is working really hard to develop helios brain which
02:23	is only a series of algorithms or steps that helios
02:26	can take to take his sensor data to control each
02:29	of his 36 joints and motors Without this brain or
02:33	these algorithms in place , helios is really just a
02:36	very expensive £330 pile of metal . The M .
02:43	I . T . Team is working hard on algorithms
02:45	that improve helios ability to walk on difficult terrain .
02:49	Currently , a human operator must send commands to Helios
02:52	through a wirelessly connected computer . The operator looks at
02:56	the sensor readouts from helios and decides what basic actions
02:59	are . All I should do . This is sort
03:01	of like how my frontal lobe took the visual input
03:04	from the back of my brain and made basic plans
03:07	for my motor system to carry out Helios operator makes
03:10	these decisions for him . But once the command is
03:13	sent to Helios , it's up to the robot to
03:15	execute them himself . So Helios brain is kind of
03:19	a hybrid . It's part human and part computer .
03:22	But why do we even need our human operator ?
03:25	Well , that's because as difficult as motor control is
03:29	. Motor planning is even harder . It takes a
03:31	lot of knowledge about your environment to translate your goals
03:34	into plants to us humans . It's really obvious how
03:38	to pick up this drill . But that's because we
03:40	already know so much about it . We know that
03:42	it's solid . You know about how much it weighs
03:45	that's being held up by the table and we know
03:47	to pick it up from the handle . But the
03:50	program all of this information and so much more into
03:53	helios , it would be impossible to do it for
03:56	every conceivable object . However , there are some labs
03:59	that are working on other robots that can learn this
04:01	information slowly from its environment , like we do when
04:04	we're infants or toddlers for helios to perform a task
04:07	completely on his own . His creators have to answer
04:10	one more question how to adapt to an ever changing
04:13	environment . If I go to pick up this drill
04:15	, but someone takes it before I can get there
04:17	I can run after them and take it right back
04:21	. Yeah . But if I were to take this
04:23	drill from helios , he wouldn't know what to do
04:25	without his human operator . It takes 30-50% of our
04:30	brain power just to keep up with all the sensory
04:32	information from our eyes . So you can imagine how
04:35	challenging it would be to program a robot to continuously
04:39	pay attention to his surroundings and adapt to them .
04:42	Fully autonomous robots are on their way . But partially
04:45	autonomous robots like helios can still be very helpful while
04:49	we're waiting . Maybe one day soon , firefighters will
04:52	be controlling robots that can carry people from burning buildings
04:55	and astronauts might even do their spacewalks remotely controlling a
04:58	robot like helios from Earth .

Summarizer

DESCRIPTION:

OVERVIEW:

Humanoid Robot Brains (Science Out Loud S1 Ep2) is a free educational video by MITK12Videos.

This page not only allows students and teachers view Humanoid Robot Brains (Science Out Loud S1 Ep2) videos but also find engaging Sample Questions, Apps, Pins, Worksheets, Books related to the following topics.