imagine we asked an engineer to build a
humanoid with two arms or say with one
arm so that it does not take so long and
actually if you ask my team what the one
thing is that I should still learn it’s
waiting I’m really not patient and of
course the engineer needs a bit more of
information so we say well of course the
robot should be able to somehow sense
where up-and-down is and probably should
be able to lift its arm or maybe like
this way better and say a third task
would be probably okay everybody is
talking about human machine interaction
so let’s say you should be able to
detect touch and maybe then react on
that say two days later the engineers
come back comes back and says well okay
here’s the system well and say well
first of all what took you so long and
the answer then is probably well you
know you want this robot to have an arm
okay that’s easily done but then of
course we need a sensor here in the arm
for the angle and then we need these
sensors in the mobile phone which detect
if it’s upside down or the other way
around acceleration sensors what else
and then of course our touch sensors
this gave me the creeps I invented here
a skin or something or maybe some
buttons and of course this took some
time so two days is pretty fast and I
said okay that’s a kind of a
sophisticated system and yes we all the
time try to build more sophisticated
systems so we put something on top a
side of it and it gets more complex from
day to day so I asked him okay but I my
idea was like something more simple
robots not so many cables so many
sensors can you do better can’t you do
better and say well I don’t have a clue
I said well watch he
I Sarah you’re giving me the finger
right that’s I’m polite so no watch here
on top of my finger sits an amoeba if
it’s a German engineer you say I need a
mover right so that’s an organism which
is consists only of one cell no neural
network nothing no different organelles
so it’s just one cell and this cell can
sense up and down it can move and so on
so maybe we can copy something from
there isn’t there way to do it easier as
I I have no clue how could we do this
and I say well okay let me draw a simple
diagram we will not go into detail here
just ignore the technological stuff I
mean it’s true this is what is in the
box but the story he is right you can
just do it it will work so we have a
power supply of course you need a
battery or something and on the other
side we have the motor and maybe some of
you have been playing with Lego in your
childhood so this is how you can build
it and there’s just in between the power
supply and the motor there’s a small box
like the amoeba and no sensors and so on
well the engineer of course says well
you must you must be bullshitting me I
mean this how could this work so you say
okay I give another try
so here’s the answer here’s the Lego
motor and there’s a bit of a plastic rod
on it and on the other side you have the
battery and this is the proof that’s
just a single cable and I mean this is
not slow-motion this is how it’s getting
up so you see the first task is soft and
it’s not in this video but you can
eventually turn it around so it will
reliably detect up and down and it will
stop there in the upright position
okay good I don’t know how you did it
but okay but what for the touch I mean
how would a piece of plastic detect
touch so here’s the answer or at least
the proof that this can work very
reliably you see if you touch the rod of
course you cannot charge something
without at least having a slight force
upon the thing and this of course can go
through the motor back through the cable
to the electronic circuit and be
processed there and the nice thing of
course is
and that’s actually the bottom stuff the
bottom line here of the story this
system is so ridiculously easy that its
utmost robust
I mean if you cut off part of this
plastic rod or it burns down or
something the rest is still able to
detect the touch I mean come on if you
put on pressure sensors or something you
need a huge amount of cabling and if
something breaks here the informations
lost and the abilities come so that’s
actually how in the lab we try to
reinvent and build thinking about robots
building robots and getting down with
complexity maybe also something which is
helpful for financial systems politics
and so on but I will focus on this very
robot here okay so far for the Lego now
you say okay that’s nice it’s a nice
gimmick but I mean how about a
full-sized robot this may be fifty
mortars and so on so the proof is
actually you can scale it up to a user
system to a larger system so Simon would
would you please be so nice so that’s a
single arm of the robot and as you see
the arm is very much like the Lego motor
able to get into an upright position I
mean you just have to believe me but you
can do this without programming there’s
no microcontroller and this Lego recipe
and you see so I touch it it goes
against me if I let go it goes up again
thank you so much so that’s possible and
the nice thing is you don’t need a
central intelligence who wants a central
intelligence anyway I mean they are
collecting data and god knows what they
do with that so it’s nice for the system
to have something which is already
working somehow it’s also a matter of a
question of how you work together in the
team in many laps you have a huge team
and one very expensive machine and they
compete so can I have the machine for
some for a while and here in our lap you
see often one guy with a head and a
torso and one arm and the other one with
the torso and two legs and no arms and
no hat to do some walking
experimentation right and in the end of
the day they gets together and
try to put it all together now you said
well this is actually one joint I mean
it’s an elbow so what and there is not
really a huge force on this and at some
point you need to drive the system how
it can behave but actually as it turns
out it’s not it’s not necessary and
sometimes you find this accidentally say
Oh watch what’s in there actually to be
honest when I built this Lego system I
was not aware of the fact that if you
touch it it gets down I just wanted the
first solution but when I touched it I
said well actually the the other problem
is soft also now let’s go for a more
sophisticated thing and we will see the
lag of the robot and here it’s three
joints so Benjamin if you if you will
thank you so much and so it’s about the
ankle the knee and the hip joint and
before you started it will go upside and
the movements of the different joints
will orchestrate itself so there is no
central processor orchestrating and
there’s also no calculated trajectory
normally robots get a command this joint
should be this angle and so on and over
time it behaves like this this is not
the case it’s just doing this like with
the Lego system
not so many talking please let go and
just watch maybe the knee joint it’s a
bit a very tiny table that I got is here
and there’s no start and stop I mean he
was just connecting the battery actually
and if you see it’s constantly waving
forward and backward like actually if
you watch yourself from time to time you
never standing a hundred percent still
and you’re also never standing
completely upright actually you’re
constantly sensing is gravitation still
there of course it is right but if we
would put this into water where it’s
flowing so there’s a squid from Tintin
fish hour which actually is floating in
the water
and if fish comes along and it touches
the tentacle it’s actually doing exactly
this so food intake and many other
behaviors thank you thank you so much
all you are that’s your applause is your
loss Benjamin yeah so this thing is
amazing isn’t it and it’s the same
principle as with illegal motor so you
can get somewhere with that and we will
in in a while see the full robot system
where the two legs arrange to have the
robot standing and the arms can then do
other thing of that now can we blame the
poor engineer of course not and why did
he not come up with the solution in the
first place and I know because I was the
engineer I said why did it took you so
long and so on so no offense here it’s
the educational system and this is also
part of of my talk actually an engineer
is educated this way if you want a sense
XYZ you need an X Y and set sensor and
nowadays everything is put into a
computer or in a microcontroller it’s
put together and so on and then the path
is already laid out and of course to
break rules you have to go away from the
mainstream and try to rethink it now
this principle of modularity can also be
ported over into the head where we have
cognitive systems like scene movement or
speech and so on and probably we can
first of all look through the eye of the
robot who never want to do this so maybe
we have a look maybe you can point it to
the audience
do we see something I know okay this is
later now we start with audio one thing
after the other so here you see my
spectrum No please let us see it we stay
here at this image so if I’m quiet which
is of course hard for me you see but all
black
II amoeba amoeba maybe you can borrow an
idea from the amoeba okay so this is how
it works and from there we get in a
simple system we can cut things out and
go into a graph like representation
about nearness so we have grow we have
show we have I don’t know there’s also
shy in the dictionary so if I say for
example my own listen shy we can have
this actuated and then there’s like an
associative Network from where we can go
on maybe say a new word like my own
listen amoeba and then there is part of
the spectrogram which can then be
introduced and another thing is of
course vision so now we can look through
so maybe you just point to the audience
for a while
so here you see left and the other way
around would be right and so on and yeah
maybe we stay here and just put those
part-time hard solutions together on the
desk and I can then maybe think loud
about the educational system so the idea
here is to change it and actually we are
preparing a bachelor’s program which
will start October next year where we
try to cross those classical engineering
thinking which is don’t get me wrong
perfect for most of the things I mean
who in a factory wants have a robot
which does what it likes to do for that
it’s of course the natural ways okay
but to get beyond those and to get a
robust system in the way that we saw you
need to really get down with complexity
throw things out and this is extremely
difficult all the time still after 10
years you have to ask yourself can I
leave still things out is this cable
necessary is this sensor necessary and
in the end often you are amazed what’s
possible but you have to cross the
border here and we try to
in a bachelor program so that they start
from the beginning then the sooner you
are on this traditional engineer track
of the harness in the end to not do the
solution yourself but to try to enable
the robot to come up with a solution
okay so last part of the show we will
put the robot now it’s as you may have
seen it’s connected there’s two arms
here one head and of course is a very
phred child child so this is nothing for
warfare it will not harm anybody right
go for it and just to give you an idea a
master student easily takes a year to go
into a standing up motion we will see
this Benjamin just do it will help the
robot to stand up and there’s fancy
stuff going on in the joints who is
dominating whom in the beginning the arm
is stiff but then when the robot found
its stable position then it’s released
again and you guessed it while Benjamin
is going to the front with the robot
that all those things rely on this
simple principle this raising up the arm
or having the robot standing is one of
the principles yeah it’s a guided
walking people said it’s more like a
very old patient right maybe because I
have been involved in the development of
the robot this could be but of course
there’s not one way of walking you can
do this in very different ways ok so we
come to stand and maybe we have the
robot wave of the hands and say goodbye
and in the end actually if you’re
interested or you know people who like
to think the other way around
you’re welcome to tell them and try to
go different ways starting next year in
winter and once it’s stable
Benjamin how okay
well if not we just take it out and we
actually are in the break he will find
us they understand you can take a selfie
with the robot and we will have the
robot waving there for you so it’s not
the fool goodbye for but for now we say
thanks and come and visit us
oh well okay see you later in the break
that’s it
[Applause]
like apnea
[Applause]
thankfully okay show isn’t over
yeah well come on you can do it yeah
[Applause]
you
humanoid with two arms or say with one
arm so that it does not take so long and
actually if you ask my team what the one
thing is that I should still learn it’s
waiting I’m really not patient and of
course the engineer needs a bit more of
information so we say well of course the
robot should be able to somehow sense
where up-and-down is and probably should
be able to lift its arm or maybe like
this way better and say a third task
would be probably okay everybody is
talking about human machine interaction
so let’s say you should be able to
detect touch and maybe then react on
that say two days later the engineers
come back comes back and says well okay
here’s the system well and say well
first of all what took you so long and
the answer then is probably well you
know you want this robot to have an arm
okay that’s easily done but then of
course we need a sensor here in the arm
for the angle and then we need these
sensors in the mobile phone which detect
if it’s upside down or the other way
around acceleration sensors what else
and then of course our touch sensors
this gave me the creeps I invented here
a skin or something or maybe some
buttons and of course this took some
time so two days is pretty fast and I
said okay that’s a kind of a
sophisticated system and yes we all the
time try to build more sophisticated
systems so we put something on top a
side of it and it gets more complex from
day to day so I asked him okay but I my
idea was like something more simple
robots not so many cables so many
sensors can you do better can’t you do
better and say well I don’t have a clue
I said well watch he
I Sarah you’re giving me the finger
right that’s I’m polite so no watch here
on top of my finger sits an amoeba if
it’s a German engineer you say I need a
mover right so that’s an organism which
is consists only of one cell no neural
network nothing no different organelles
so it’s just one cell and this cell can
sense up and down it can move and so on
so maybe we can copy something from
there isn’t there way to do it easier as
I I have no clue how could we do this
and I say well okay let me draw a simple
diagram we will not go into detail here
just ignore the technological stuff I
mean it’s true this is what is in the
box but the story he is right you can
just do it it will work so we have a
power supply of course you need a
battery or something and on the other
side we have the motor and maybe some of
you have been playing with Lego in your
childhood so this is how you can build
it and there’s just in between the power
supply and the motor there’s a small box
like the amoeba and no sensors and so on
well the engineer of course says well
you must you must be bullshitting me I
mean this how could this work so you say
okay I give another try
so here’s the answer here’s the Lego
motor and there’s a bit of a plastic rod
on it and on the other side you have the
battery and this is the proof that’s
just a single cable and I mean this is
not slow-motion this is how it’s getting
up so you see the first task is soft and
it’s not in this video but you can
eventually turn it around so it will
reliably detect up and down and it will
stop there in the upright position
okay good I don’t know how you did it
but okay but what for the touch I mean
how would a piece of plastic detect
touch so here’s the answer or at least
the proof that this can work very
reliably you see if you touch the rod of
course you cannot charge something
without at least having a slight force
upon the thing and this of course can go
through the motor back through the cable
to the electronic circuit and be
processed there and the nice thing of
course is
and that’s actually the bottom stuff the
bottom line here of the story this
system is so ridiculously easy that its
utmost robust
I mean if you cut off part of this
plastic rod or it burns down or
something the rest is still able to
detect the touch I mean come on if you
put on pressure sensors or something you
need a huge amount of cabling and if
something breaks here the informations
lost and the abilities come so that’s
actually how in the lab we try to
reinvent and build thinking about robots
building robots and getting down with
complexity maybe also something which is
helpful for financial systems politics
and so on but I will focus on this very
robot here okay so far for the Lego now
you say okay that’s nice it’s a nice
gimmick but I mean how about a
full-sized robot this may be fifty
mortars and so on so the proof is
actually you can scale it up to a user
system to a larger system so Simon would
would you please be so nice so that’s a
single arm of the robot and as you see
the arm is very much like the Lego motor
able to get into an upright position I
mean you just have to believe me but you
can do this without programming there’s
no microcontroller and this Lego recipe
and you see so I touch it it goes
against me if I let go it goes up again
thank you so much so that’s possible and
the nice thing is you don’t need a
central intelligence who wants a central
intelligence anyway I mean they are
collecting data and god knows what they
do with that so it’s nice for the system
to have something which is already
working somehow it’s also a matter of a
question of how you work together in the
team in many laps you have a huge team
and one very expensive machine and they
compete so can I have the machine for
some for a while and here in our lap you
see often one guy with a head and a
torso and one arm and the other one with
the torso and two legs and no arms and
no hat to do some walking
experimentation right and in the end of
the day they gets together and
try to put it all together now you said
well this is actually one joint I mean
it’s an elbow so what and there is not
really a huge force on this and at some
point you need to drive the system how
it can behave but actually as it turns
out it’s not it’s not necessary and
sometimes you find this accidentally say
Oh watch what’s in there actually to be
honest when I built this Lego system I
was not aware of the fact that if you
touch it it gets down I just wanted the
first solution but when I touched it I
said well actually the the other problem
is soft also now let’s go for a more
sophisticated thing and we will see the
lag of the robot and here it’s three
joints so Benjamin if you if you will
thank you so much and so it’s about the
ankle the knee and the hip joint and
before you started it will go upside and
the movements of the different joints
will orchestrate itself so there is no
central processor orchestrating and
there’s also no calculated trajectory
normally robots get a command this joint
should be this angle and so on and over
time it behaves like this this is not
the case it’s just doing this like with
the Lego system
not so many talking please let go and
just watch maybe the knee joint it’s a
bit a very tiny table that I got is here
and there’s no start and stop I mean he
was just connecting the battery actually
and if you see it’s constantly waving
forward and backward like actually if
you watch yourself from time to time you
never standing a hundred percent still
and you’re also never standing
completely upright actually you’re
constantly sensing is gravitation still
there of course it is right but if we
would put this into water where it’s
flowing so there’s a squid from Tintin
fish hour which actually is floating in
the water
and if fish comes along and it touches
the tentacle it’s actually doing exactly
this so food intake and many other
behaviors thank you thank you so much
all you are that’s your applause is your
loss Benjamin yeah so this thing is
amazing isn’t it and it’s the same
principle as with illegal motor so you
can get somewhere with that and we will
in in a while see the full robot system
where the two legs arrange to have the
robot standing and the arms can then do
other thing of that now can we blame the
poor engineer of course not and why did
he not come up with the solution in the
first place and I know because I was the
engineer I said why did it took you so
long and so on so no offense here it’s
the educational system and this is also
part of of my talk actually an engineer
is educated this way if you want a sense
XYZ you need an X Y and set sensor and
nowadays everything is put into a
computer or in a microcontroller it’s
put together and so on and then the path
is already laid out and of course to
break rules you have to go away from the
mainstream and try to rethink it now
this principle of modularity can also be
ported over into the head where we have
cognitive systems like scene movement or
speech and so on and probably we can
first of all look through the eye of the
robot who never want to do this so maybe
we have a look maybe you can point it to
the audience
do we see something I know okay this is
later now we start with audio one thing
after the other so here you see my
spectrum No please let us see it we stay
here at this image so if I’m quiet which
is of course hard for me you see but all
black
II amoeba amoeba maybe you can borrow an
idea from the amoeba okay so this is how
it works and from there we get in a
simple system we can cut things out and
go into a graph like representation
about nearness so we have grow we have
show we have I don’t know there’s also
shy in the dictionary so if I say for
example my own listen shy we can have
this actuated and then there’s like an
associative Network from where we can go
on maybe say a new word like my own
listen amoeba and then there is part of
the spectrogram which can then be
introduced and another thing is of
course vision so now we can look through
so maybe you just point to the audience
for a while
so here you see left and the other way
around would be right and so on and yeah
maybe we stay here and just put those
part-time hard solutions together on the
desk and I can then maybe think loud
about the educational system so the idea
here is to change it and actually we are
preparing a bachelor’s program which
will start October next year where we
try to cross those classical engineering
thinking which is don’t get me wrong
perfect for most of the things I mean
who in a factory wants have a robot
which does what it likes to do for that
it’s of course the natural ways okay
but to get beyond those and to get a
robust system in the way that we saw you
need to really get down with complexity
throw things out and this is extremely
difficult all the time still after 10
years you have to ask yourself can I
leave still things out is this cable
necessary is this sensor necessary and
in the end often you are amazed what’s
possible but you have to cross the
border here and we try to
in a bachelor program so that they start
from the beginning then the sooner you
are on this traditional engineer track
of the harness in the end to not do the
solution yourself but to try to enable
the robot to come up with a solution
okay so last part of the show we will
put the robot now it’s as you may have
seen it’s connected there’s two arms
here one head and of course is a very
phred child child so this is nothing for
warfare it will not harm anybody right
go for it and just to give you an idea a
master student easily takes a year to go
into a standing up motion we will see
this Benjamin just do it will help the
robot to stand up and there’s fancy
stuff going on in the joints who is
dominating whom in the beginning the arm
is stiff but then when the robot found
its stable position then it’s released
again and you guessed it while Benjamin
is going to the front with the robot
that all those things rely on this
simple principle this raising up the arm
or having the robot standing is one of
the principles yeah it’s a guided
walking people said it’s more like a
very old patient right maybe because I
have been involved in the development of
the robot this could be but of course
there’s not one way of walking you can
do this in very different ways ok so we
come to stand and maybe we have the
robot wave of the hands and say goodbye
and in the end actually if you’re
interested or you know people who like
to think the other way around
you’re welcome to tell them and try to
go different ways starting next year in
winter and once it’s stable
Benjamin how okay
well if not we just take it out and we
actually are in the break he will find
us they understand you can take a selfie
with the robot and we will have the
robot waving there for you so it’s not
the fool goodbye for but for now we say
thanks and come and visit us
oh well okay see you later in the break
that’s it
[Applause]
like apnea
[Applause]
thankfully okay show isn’t over
yeah well come on you can do it yeah
[Applause]
you
Please follow and like us: