►
Description
Part of the mandate to the AI/ML working group is to do a prototype to test AI/ML techniques of interest. Using ultrasonic sensors might be an inexpensive path forward. Here is what a possible testbed might look like.
A
Hey
everybody,
so
here's
an
sr04,
echo
location
array:
these
are
ultrasonic
transducers,
they
send
out
a
ping
and
then
you
can
program
a
controller
to
count
the
time
it
takes
for
the
Ping
to
go
out
and
then
come
back.
There's
a
keep
out
to
eliminate
ring
down.
So
you
you
can't
get
arbitrarily
close
a
couple
of
centimeters
anyway.
This
is
a
an
array
built
for
a
robot
based
on
the
Zooka.
A
So
Zooka
has
some
Motors
here
they
are
and
some
some
batteries
here
they
are
and
a
motor
controller
and
over
here
is
Arduino
that
that
runs
the
the
motor
controller,
okay,
but
the
sensing
for
for
moving
around
objects.
What
we
came
up
with
was
an
array
sort
of
on
an
arc.
You
can
see
it's
a
laser
cut,
acrylic
Arc
and
here's
all
the
the
sensors
okay.
So
you
should
probably
recognize
these
are
ultrasonic
transducers.
A
A
We
talked
about
using
these
this
type
of
sensor,
acoustic
or
Ultrasonics
to
try
out
ideas
from
Dynamic
Spectrum
sensing.
It's
cheap
and
easy.
These
are.
These
are
really
inexpensive
devices
and
you
don't
need
a
powerful
computer
to
run
them
so
I'm
thinking
that
that
maybe
this
might
be
a
good
test
bed.
A
So
this
this
looks
like
a
good
idea
anyway,
we'll
we'll
see
what
we
can
do,
since
we
already
were
kind
of
like
going
in
this
direction
anyway,
for
robotics
and
I
should
probably
say
that
that
you
know
the
Echolocation
has
some
drawbacks.
So
anybody
that's
actually
used
these
no's
that
they
behave
in
certain
ways,
and
so
you
can
have
a
sensor
array
here
and
even
get
some
resolution
Improvement
by
by
having
a
you
know
a
number
of
of
beams.
A
So
who
knows,
there's
maybe
a
cool
prototype
to
to
explore
Dynamic
spectrum
allocation
and
try
out
algorithms
in
a
way,
that's
a
whole
lot,
cheaper
and
easier
than
building
microwave
test
beds
or
trying
to
do
this
over
the
air
with
existing
services.