DPRG mailing list archives: [DPRG] Cool Christmas Presents

Wed Dec 27 22:20:04 CST 2006

Ed O.,

I have a placed pictures of the disassembly process (more or less) up here:
http://edparadis.com/radar/

I also educated myself some on the aspects of radar via Wikipedia:
http://en.wikipedia.org/wiki/Doppler_effect
http://en.wikipedia.org/wiki/Doppler_radar

What we appear to have out of the box is a 'continuous wave radar':

http://en.wikipedia.org/wiki/Continuous_wave_radar

The upside being that it is simple, the downside being that it can
only detect moving objects.  I'm not convinced that is too bad of a
thing, since a robot generally moving around a lot.  This means that
everything else that stands still will appear as a moving object.
Just detecting it isn't immediately useful.

That page mentions a type of radar called "FM-CW" for "frequency
modulated continuous wave".  You ramp the transmitted frequency up and
down at a known rate, and then use that to determine range.  The page
describes it pretty well.  The Austech forum mentioned that you could
tune some oscillators by changing their supply voltage.  I think this
only applies to gun-diode type oscillators, however.  We may be able
to subtly change the frequency of our transmitter by simply modulating
the supply voltage.  It appears a promising route.  The oscillator is
a single transistor oscillator, and I know you can tune single
transistor VHF oscillators to a degree using supply voltage changes.

Questions to answer:
1) can you change it
2) how do you know you've changed it
3) since you'll be looking at time of flight of a radio signal, can
you even measure that in the first place?  That is, sure you can
change the signal being transmitted, but can you detect that change
quickly enough to have a useful range?

You can also just pulse the transmitter:

http://en.wikipedia.org/wiki/Pulse-doppler_radar

That system looks at the phase differences of incoming pulses.  Phase
changes in a hetrodyning system would show up as amplitude changes (as
do frequency changes, since they're essentially phase changes).  Due
to this situation, you could probably measure distance or velocity,
but not both.  You would have to know velocity is zero to measure
distance.  Distance technically makes a difference for measuring
velocity, and I think looking at the amplitude envelope of the output
of the mixer could tell you something about distance.

You could also just look at the amplitude of the mixer output and time
the difference between the start of transmission and the peak of the
return.  This is awfully quick, I would expect.  This is in essence
the same as most of the available sonar units, but using radio waves.

Also, I did some math to figure out the shift.  Assuming you are
standing still, and the object you are measuring is moving, the
frequency shift you will see is twice the typical frequency shift of
the classic "Doppler equation".  This is because you see one shift
when you transmit to the moving target, and then another shift when
the moving target bounces (transmits) that signal back.  Since the gun
measures the difference between the TX and RX frequency, you can do
some algebra and find:

Fo = Ftx(1 - (c+vm) / (c-vm) )
where
Fo is the frequency of the output on the white wire of the waveguide module
Ftx is the transmitting frequency (10.525 GHz)
c is the speed of light (299.79e6 m/s)
vm is the velocity of the moving target in m/s

If vm = 1 m/s, then Fo = 70.2 Hz.  (We ignore the sign of the result
because of the mixer).  For vm = 44.7 m/s (100 mph), Fo = 3.137 kHz.

Ed P.

On 12/27/06, Ed Okerson <ed at okerson.com> wrote:
> Ed P.,
>
> Found that site about 30 seconds after my e-mail was sent. :)
>
> 0-100Hz sounds about right for Doppler shift.  Hook it to an audio amp and
> walk across the room in front of it, the tone should change with your
> speed.
>
> Now I wonder how to convert them to Time-Of-Flight?  The Doppler might be
> useful as an alternate way of getting ground speed from a large outdoor
> robot, or measuring the speed of quick trip robots, but not much else.  A
> true TOF unit would be a really cool alternative to sonar and IR though,
> with a much larger potential range.
>
> Ed O.
>
> > Ed O.,
> >
> > I just read the entire thread here:
> > http://www.austech.info/showthread.php?t=135840&page=1
> >
> > They seem to be most concerned with using the measuring electronics
> > with other more powerful TX/RX pairs to make homebrew radar guns, but
> > the posts by "mitaux8030" and "Trash" are pretty informative.
> >
> > I just finished tearing the beast apart, etc.  First off: its a real
> > beast to get apart.  You have to drill out the plastic plugs to get at
> > the screws.  A big hint is to get the two end caps off first.
> >
> > I've got the waveguide/TX/RX/mixer unit on my bench.  The signal
> > coming out of it doesn't seem to get any better by powering the unit
> > continuously versus the 25% on, 8khz duty cycle the unit originally
> > uses.
> >
> > I need to build an amp and fiddle with some filters to see if I can
> > get an obviously measurable signal out of it.  I'm also need to find a
> > book to figure out what sort of frequency range to expect out.
> > 0-100Hz seems to appear, but I'm not sure if thats due to real doppler
> > affect, or my hand affecting the near-field of the TX/RX pair.
> >
> > I've got pictures of the gun disassembled. They sort of fill the gaps
> > of the pictures on that forum.  I'll get them up on a site soon.
> >
> > Ed P.
> >
> > On 12/27/06, Ed Okerson <ed at okerson.com> wrote:
> >> Ed,
> >>
> >> What search terms did you use?  I have not been able to find any useful
> >> info so far.
> >>
> >> Ed
> >>
> >> > Ed,
> >> >
> >> > When I got back to Austin today, I stopped by the local Walmart and
> >> > picked up one of those hot wheels radar guns.  Pretty slick!  They
> >> > appear to be on clearance for $20.00 and I saw them at HEB later for
> >> > $12.50.
> >> >
> >> > I found some pictures online, and it seems that its using doppler
> >> > shift, not time of flight.  Makes sense, because thats cheap.   There
> >> > is some discussion on boosting the power, too.
> >> >
> >> > But it seems that the base signal is being measured and displayed by a
> >> > microcontroller of some sort.
> >> >
> >> > So the question to everyone is this:
> >> > Given a sensor that gives you the speed of a moving object, can you
> >> > produce 'interesting' data for navigation?
> >> >
> >> > Perhaps use it while driving to check your odometry by measuring your
> >> > own speed against the ground?
> >> >
> >> > Sweep it 360 degrees and create a sort of optical flow map?
> >> >
> >> > Use two at once and come up with the speed of two points in front of
> >> > you, then compute a line (gradient) that could give you speed and
> >> > direction of travel?
> >> >
> >> > I've not opened mine up yet, but from the pictures, the base-band
> >> > processor would easily be replaced.
> >> >
> >> > Ed P
> >> >
> >> >
> >> > On 12/26/06, Ed Okerson <ed at okerson.com> wrote:
> >> >> So what cool toys did everyone find under the tree?
> >> >>
> >> >> My kids got me the Hot Wheels Radar Gun.  How they manage to make a
> >> real
> >> >> Doppler radar gun for $20 is beyond me.  Now to figure out a good
> >> >> robotics
> >> >> use for it.  I wonder how hard it would be to make a sonar
> >> replacement?
> >> >>
> >> >> Anyone going to RBNO tonight?
> >> >>
> >> >> Ed
> >> >>
> >> >> _______________________________________________
> >> >> DPRGlist mailing list
> >> >> DPRGlist at dprg.org
> >> >> http://list.dprg.org/mailman/listinfo/dprglist
> >> >>
> >> >
> >>
> >>
> >>
> >
>
>
>