►
From YouTube: Open Source Robotic Tractor - Meeting 4
Description
Working document is still at http://opensourceecology.org/wiki/Open_Source_Robotic_Tractor
-----
What you see here at Open Source Ecology is an ambitious program based on a volunteer effort. To help us reach the goals - please consider joining as an OSE Developer in 2017-
http://opensourceecology.org/wiki/OSE_Developers
Take a minute to subscribe to our email newsletter (updates, workshops, etc): http://bit.ly/1LtcM44
B
B
Now,
from
anytime,
you
come
back
until
we
will
see
what
we're
doing
with
it,
but
so
you
get
this
three-dimensional
feed
from
it.
The
visuals
and
I
said
if
it's
actually
use
Arbutus
file
to
generate
this
dynamically,
so
NER
this
file,
you
uploaded
we're
able
to
remember
it
pretty
much
and
then
what
you're
seeing
here
is
the
black
and
white
floor
tiles.
Here,
that's
kind
of
coming
from
gee
mapping
for
feeling
about
floorplan.
C
B
And
the
things
that
you
see
coming
up
or
the
color
and
RGB
data
from
a
blank
tile
sensor
and
and
that
data
is
being
collected
using
our
it's
called
our
tab
map,
which
is
a
slam,
mapping
tool
and
the
RRS
ecosystem.
And
then
you
can
see
the
driving
that
we
actually
have
about
driving
and
those
red
arrows
are
kind
of
I,
think
50
or
so
odometry
points.
And
then
the
pink
lines
you
see
were
the
plans
that
it
had.
B
While
it's
driving
so
anytime,
we
drop
in
a
new
marker
to
go
toward
it'll
redraw
that
people
pink
long,
and
so
you
can
see
all
these
different
topics
of
data
flying
around
and
we
can
actually
credentialing
your
user
out
and
based
off
of
all
the
topics
and
devices
a
beauty
mark,
and
so
we're
actually
also
adding
in
into
end
cryptography.
So
if
you
want
to
be
able
to
have
a
high
degree
of
security,
so
you
know
that
you
know
that
we
can't
even
observe
the
content
of
your
messages.
B
B
Resource
play
yeah
so
that
all
of
the
parts
that
are
required
to
do
get
your
robot
activated
or
open
source
and
then
all
the
libraries
you
need
to
get
your
data
in
and
out
of
the
cloud
or
of
the
source.
We
unfortunately
had
to
do
a
lot
of
proprietary
in
the
cloud
to
be
able
to
scale
the
solution.
To
you
know
hundreds
of
thousands
of
users
so
we're
not
able
to
open
source
a
lot
of
that
work.
B
B
B
But
we're
actually
using
this
robot
from
a
called
you've,
been
Queenie
robotics,
mm-hmm,
they're,
building
it's
actually
about
the
same
price
range
as
a
turbot
to
and
yeah
so
where
I
started
without
over.
Here
we
have
this
front
desk
in
our
lobby
here,
and
so
we
were
driving
down
through
the
lobby.
That's
a
robots
come
to
left
over
here
by
the
refrigerator.
B
That's
the
ability
team
goal.
They
gave
us
early
prototype,
and
so
it's
it's
actually
a
pretty
good
little
worker
purse.
It
runs
for
about
eight
hours
or
so
mmm-hmm
on
a
single
charge
and
then
I'm
and
then
once
it's
all
charged
up,
you
can
actually
load
it
up
with
quite
a
bit
of
weight
for
such
a
small
platform.
It'll
actually
carry
about
a
hundred
kilograms.
B
C
B
Additional
odometry
different
from
that
for
very
cheap
and
then
more
expensive
side.
We
just
kind
of
slashed
on
a
joystick
and
of
another
Raspberry
Pi
back
here,
and
this
is
a
point
cloud
sensor
and
then
up
here's
a
scans.
We
light
our
unit
so
actually
right
now,
when
I
was
driving
around
we're
only
using
the
point
cloud
sensor
and
then
just
odometry
from
the
tires.
So
that
was
big
stitch
just
for
the
tire
odometry
that.
B
E
B
D
B
B
If
they
actually
picked
up
two
infrared
cameras
and
they
actually
do
a
stereo
infrared
matching,
and
then
we,
if
you
go
indoors
with
it
at
all
or
if
it
gets
dark
enough,
it'll
switch
to
the
projected
pattern
and
they're
only
about
$200,
depending
on
which
stuffs
are
you
looking
at
mm-hmm
we
play
around
a
bit
with
those
another
option
for
outdoor
use.
Is
the
so
the
nice
thing
about
honey
cloud
sensors?
Is
they
don't
require
a
huge
amount
of
processing
power
to
get
you
the
3d
data?
Of
course?
B
A
D
B
I
mean
all
of
this:
that's
just
actually
the
so
XT
on
a
sadly
their
business
got
acquired
by
Apple
or
something
like
that
many
years
ago.
So
they
stopped
selling
sensor.
But
it
actually
probably
a
great
sensor
if
you
want
the
down
facing
camera,
that
looked
at
your
plants
and
go
through
the
actually
prefer,
because
that,
because
it's
probably
with
a
bit
from
the
sunlight
yeah.
D
Excellent,
let's
see
Matt
can
you
fill
me
in
on
as
far
as
Alan
Alan
are
you
are
you
you
asked
Alan
to
to
showcase
some
of
the
potential
sensing
hardware
or
tell
me
more.
B
A
D
D
F
A
D
D
E
A
Driving
around
yeah
then
Tuesday.
We
can
start
installing
some
of
the
electronics
and
maybe
to
characterize
the
vehicle.
How
does
it
drive?
What's
the
response
Thursday?
Are
we
going
to
st.
Louis
for
down
there
and
then
Friday
journey
I'll
come
back
up
to
load
software?
Do
the
testing
and
restoration
and.
D
A
F
D
A
A
D
Far
as
the
electronics
that
just
the
basics
is
just
getting,
you
know
taking
out
the
solenoids
and
making
sure
that
you
can
activate
them
with
the
you
know,
with
the
software
that's
step,
one
yeah
good
go
go
with
that,
but
yeah
that
actually
works
well.
So
because
we
were
talking
about
using
one
of
our
older
tractors,
but
we
actually
are
want
to
strip
its
tracks.
We're
gonna,
reuse,
the
tracks
of
that
and
because
we're
building
the
two
tractors,
one
is
the
tiny
one.
D
F
So
we
started
the
end
started
this
sound
kind
of
I,
don't
know
if
the
github
Roxanna
AG
page,
but
this
where
Ian
furnished
and
everything
it
looks
like
parts
that
are
missing
there
are
for
the
simulation
we
haven't
completed
that
so
there's
not
a
little
plugin.
Yes,
I
was
going
to
start
working
on
that.
You
have
had
the
Mike.
C
F
B
The
razor
I
am
you
from
Sparkfun
is
really
well
supported.
Last
I
checked,
let
me
see
if
it's
still
being
sold
or
the
dumper
conscious
of
it.
Okay,
but
it's
a
I
believe
it's
a
nine
degree
of
freedom,
so
accelerometers
gyroscopes
and
magnetometers,
let's
see
I
see,
one
of
them
is
retired.
Now,
let's
active,
that's
where
you
put
this.
B
It's
a
nice
thing
about
the
older
razor
at
least
was
that
it
happens
to
the
possibly
on
board
on
an
atmega.
It's
off.
He
had
to
do
is
flash
up
an
arduino
sketch
on
to
do
it
and
then
Gohan
Rose
there's
just
a
driver
for
it,
and
then
there's
a
slightly
involved
tuning
process,
which
would
be
kind
of
hard
on
a
factor
to
characterize
it.
But
when
it's
free-floating,
it's
pretty
easy,
you
have
to
you,
just
have
to
rotate
it
a
bunch
to
get
it
to
be
comfortable
with
whatever
magnetic
field.
C
C
B
B
D
C
F
F
B
B
It
this
package
allows
you
to
set
the
margin
of
error,
essentially
on
every
axis,
and
then
you
can
also
tell
it,
which
axis
took
me
the
most
Q
sub
out
of
the
sensor.
Thank
you
very
much
Ross.
Thank
you.
Yeah
there's
a
Rasta
package.
It's
maintained
by
clear
path
from
X
and
it
will
it'll
do
since
refused
and
I
think
it's
Paul
River.
It's
a
robot.
Now.
B
B
Of
the
package-
and
it
kind
of
depends
us,
it
is
there's
like
five
different
packages
out
there
that
do
different
parts
of
the
Kalman
filtering
and
back
parts.
So,
okay,
yeah
I,
think
it's
robot
localization,
which
includes,
should
include
the
various
data
yeah.
So
it
includes
BKF,
localization
and
UK,
which,
depending
on
how
you
set
things
up,
is
a
good
approach,
and
so
the
cool
thing
about
this
stack
is
that
you
can
have
believed
an
arbitrary
number.
B
F
C
B
B
D
D
B
B
And
then,
once
you
have
a
command
velocity
system
that
works,
we
can
then
start
plugging
into
the
various
navigation
tools
and
the
first
cleaning
of
the
signal
I
think
that
ubiquity
they
did
it
really
well.
They
they
make
this
tool
called
blue
basic
that
doesn't
require
nearly
as
much
configuration
as
as
mu
Bais
does.
So
once
you,
of
course,
once
you
have
man
velocity
working
and
your
sensor,
fusion
go
man,
then
you
can
start
navigating
based
off
of
a
path
than
doing
leap
wants,
but
mu
Bais
requires
a
really
detailed
map
as
well.
B
B
I
think
it
should
be
in
the
packages
they
have
a
couple
other
like
they
have
a
lot
of
Raspberry
Pi
support
packages
for
like
the
rasp,
I
cam
and
things,
but
their
nav
stack
is
really
clean
and
very
simple.
So
you
don't
have
to
do
as
much
configuration
on
digging
on
it.
Please
they
get
it
to
just
generate
the
command
velocities
and
then
it's
interchangeable
with
the
more
complicated
move
base.
B
B
B
D
D
F
F
D
E
E
E
D
D
E
E
E
E
D
D
G
D
Yeah
very
nice
yeah
that
just
just
like
in
the
future
work.
Just
with
respect
to
that
I
mean
it
seems
that
the
combination
of
drones,
which
are
available
open
source
technology
with
the
open
source
tractor
that
we're
developing
I,
mean
imagine
that
your
drones
scan
your
field
like,
for
example,
say
you
grow
in
something
like
cabbage
and
or
whatever
it.
It
is
like
squash,
and
you
actually
want
to
detect
bugs
or
pests
that
you
can
see
and
detect
by
computer
vision.
D
Then
you
say:
oh
okay,
I
need
some
bacillus
thuringiensis
to
spray
them,
which
is
a
natural
herbicide
or
whatever
remedy
you
have
to
take
it'll
be
excellent.
I
mean
this
is
not
so
far
of
a
cry
to
to
do
once
we
want
to
get
these
systems
developed,
that's
which
is
very
interesting
in
terms
of
automated
agriculture,
which
could
then
the
way
I
look
at
it
is
distributed
agriculture
that
this
is
not
no
longer
relying
on
mega
farms,
but
small.
Automated
operations
can
now
compete
with
the
Giants
because
they're
just
efficient.
G
E
D
E
F
E
Will
find
it
then,
like
between
the
amounts,
the
crop
itself,
so
what
they're
planning
to
do?
The
next
is
to
figure
out
some
way
to
be.
We
actually
are
working
on
it
right
now,
a
PhD
student
who
is
senior
to
me.
He
is
working
on
the
detection,
part
of
it,
so
we're
going
to
have
the
ability
to
use
and
some
sort
of
fun
you're,
not
good
to
like
capture
to
recognize
three
different
kinds
of
beads
yeah,
and
then
our
robot
would
actually
have
a
functionality
to
display
the
display
beside.
F
E
F
E
Five
different
classes
to
five
different
classes
to
distinguish
from
so
it's
kind
of
a
tough
task
when
we
reply,
so
our
next
step
in
this
project
is
to
will
we
have
access
to
a
greenhouse
wrong
like
recently
like
just
two
days
that
being
God
access
to
the
greenhouse?
We
have
a
penis
on
gatha,
so
we're
planning
to
grow
our
own
we'd,
probably,
and
then
take
like
multiple
pictures
and
figure
out
that
data
it
works
because
the
last
time
we
tried
to
do
it
with
our
cascade
and
reserves
were
like.
D
D
A
A
E
E
D
D
Hey
hey
guys,
I
was
supposed
to
talk
to
someone
else
on
this
same
channel
I'm,
but
I
can't
tell
the
person
the
shower.
Are
you
on
here
and
if
you.