--- In sensorforum@yahoogroups.com, Doug Niessen <k6sts@...> wrote:
>
> Doug,
>
> What part of the world are you in?
>
> I'm located in CA near Sacramento and have a Seabotix
> LBV150. I have some future plans that may call for a
> positioning system like what your speaking of.  Is
> your Lampray a product?
>
> Cheers,
> Doug
Doug,

I am on Cape Cod Mass.  Imetrix, the company I worked for when we
designed the Lampray went bankrupt about three years ago.  Now two
Lamprays exist, both owned by Seaward Marine Services of Norfolk
VA,(www.seaward-marine.com).  Imetrix was trying to sell them for
about $150,000.  The positioning system we developed is called SNAP,
an offshoot of SHARPS which was developed at WHOI for use with the
Alvin submarine.  There are five SNAP systems scattered about the
globe, and I am working on improving them for Seaward's use.  SNAP
sold for $35,000 to $50,000, but I don't think Seaward wants to sell
it at all.

What is your application?  Maybe I can direct you to some of our
competitors that are still in business.

If I get this optical tracker working you will probably hear about it
here!

SherpaDoug

Doug,

What part of the world are you in?

I'm located in CA near Sacramento and have a Seabotix
LBV150. I have some future plans that may call for a
positioning system like what your speaking of.  Is
your Lampray a product?

Cheers,
Doug
--- Douglas Butler <sherpadoug@...> wrote:

> --- In sensorforum@yahoogroups.com, "Nick Alexeev"
> <kender_a@...> wrote:
> >
> > With an optical sensor, do you have to make sure
> that your rover
> > doesn't lift the silt from the bottom?
> Alternatively (and obviously)
> > you could use non-driven wheels (if you have them)
> for dead reckoning.
> >
> > Cheers,
> > Nick
> >
> With free swimming underwater robotic vehicles there
> is something
> called "dustout" when you kick up seidment which
> blinds the cameras,
> and therefore usually the pilot.  You then must wait
> a long time till
> the water is clear enough to get something done.
>
> This optical sensor is for the Lampray crawler.  It
> sticks to a ship
> hull by suction inside a rubber skirt, kind of like
> a hovercraft in
> reverse.  Spacing and propulsion is by two motor
> driven wheels and a
> caster.  If it was a ball caster one could track the
> caster movement
> for dead reckoning.  But we need a large radius (6")
> so ball casters
> are not readily available.  We would need to track
> wheel angle as well
> as rotation...sounds complicated.
>
> But the rubber suction skirt wipes most of the soft
> slime off of the
> hull so if two sensors are inside the skirt they
> should have
> reasonable visibility.  Hulls are very curved and
> have all sorts of
> nasty projections sticking out of them so I want to
> have a good
> stand-off distance and as wide a depth of field as
> possible.
>
> Our primary navigation is long baseline sonar which
> gives us position
> fixes better than +/-2cm most of the time.  But in
> around the screws
> and rudders the acoustics have problems so I am
> working on this to
> fill in the gaps.
>
> SherpaDoug
>
>


Doug  K6STS



________________________________________________________________________________\
____
Do you Yahoo!?
Everyone is raving about the all-new Yahoo! Mail beta.
http://new.mail.yahoo.com

--- In sensorforum@yahoogroups.com, "Nick Alexeev" <kender_a@...> wrote:
>
> With an optical sensor, do you have to make sure that your rover
> doesn't lift the silt from the bottom?  Alternatively (and obviously)
> you could use non-driven wheels (if you have them) for dead reckoning.
>
> Cheers,
> Nick
>
With free swimming underwater robotic vehicles there is something
called "dustout" when you kick up seidment which blinds the cameras,
and therefore usually the pilot.  You then must wait a long time till
the water is clear enough to get something done.

This optical sensor is for the Lampray crawler.  It sticks to a ship
hull by suction inside a rubber skirt, kind of like a hovercraft in
reverse.  Spacing and propulsion is by two motor driven wheels and a
caster.  If it was a ball caster one could track the caster movement
for dead reckoning.  But we need a large radius (6") so ball casters
are not readily available.  We would need to track wheel angle as well
as rotation...sounds complicated.

But the rubber suction skirt wipes most of the soft slime off of the
hull so if two sensors are inside the skirt they should have
reasonable visibility.  Hulls are very curved and have all sorts of
nasty projections sticking out of them so I want to have a good
stand-off distance and as wide a depth of field as possible.

Our primary navigation is long baseline sonar which gives us position
fixes better than +/-2cm most of the time.  But in around the screws
and rudders the acoustics have problems so I am working on this to
fill in the gaps.

SherpaDoug

With an optical sensor, do you have to make sure that your rover
doesn't lift the silt from the bottom?  Alternatively (and obviously)
you could use non-driven wheels (if you have them) for dead reckoning.

Cheers,
Nick

--- In sensorforum@yahoogroups.com, "sherpadoug" <sherpadoug@...>
wrote:
>
> Hi, I do a lot of commercial marine robotics and sensors.  I am
> looking for a better way for a robotic vehicle to track its position
> when its wheels are slipping.  I am thinking of using the optical
> sensor chip from an optical mouse (ADNS-30XX series?) but I need a
> couple of inches of ground clearance.  Has anyone used these chips
> with relay optics so they can track movement at a distance?  Are
there
> any Gotchyas to look out for?
>
> Thanks,
> SherpaDoug
>

--- In sensorforum@yahoogroups.com, "Nick Alexeev" <kender_a@...> wrote:
>
> When you get an image, what algorithm are you goind to use to calculate
> the displacement?  2D correlation?
>
That is all handled internal to the sensor chip.  It has some sort of
specialized DSP core so the chip outputs X and Y displacements as well
as a light level so you can adjust illumination and a count of
features that the chip can see so you can adjust focus.  There is a
diagnostic mode that will output the image pixel by pixel but the chip
has to slow down to do that.

SherpaDoug

When you get an image, what algorithm are you goind to use to calculate
the displacement?  2D correlation?

--- In sensorforum@yahoogroups.com, "sherpadoug" <sherpadoug@...> wrote:
>
> Hi, I do a lot of commercial marine robotics and sensors.  I am
> looking for a better way for a robotic vehicle to track its position
> when its wheels are slipping.  I am thinking of using the optical
> sensor chip from an optical mouse (ADNS-30XX series?) but I need a
> couple of inches of ground clearance.  Has anyone used these chips
> with relay optics so they can track movement at a distance?  Are there
> any Gotchyas to look out for?
>
> Thanks,
> SherpaDoug
>

I have the uM-FPU V3 running on a BX-24p via fast SPI.

Cam Thompson (of Micromega) has been very generous with his time
helping me get a fundamental understanding of the FPU, which is now
decoding real time NMEA into four floats (date, time, lat and long) in
96 bytes of FPU code.  Nice.

> ... I thought I would start with the LED based chips.

Find May, 2004 Nuts and Volts.


Tom

--- In sensorforum@yahoogroups.com, Tom Becker <gtbecker@...> wrote:
>
>  > ... can supply enough reflected signal for the array to pick it up
> over a longer distance...
>
> Perhaps, but focus is critical and the depth-of-field of the HP lens
> that is supplied with the sensor is very narrow (+/-0.2mm).
>
> The laser provides better resolution than an LED, not more range; you
> will need to illuminate the target area on your own if you do not use
> the supplied optics, too.
>
>
> Tom
>

I have quartz halogen lights for video on the vehicle.  I just have to
get the grazing angle right for the tracker.  Do the laser chips need
laser light to work?  Do they use interference patterns?  I thought I
would start with the LED based chips.

I am thinking I would use the bare chip with my own lenses.  A
standard microscope objective has a back focus of 160mm so if I use it
backwards that gives me a distance of about 6.3" which is a little too
long.  I will need some un-magnification to get more depth of field
for rough ground and also for faster ground speed.  It also needs to
be weathertight.

SherpaDoug

> ... can supply enough reflected signal for the array to pick it up
over a longer distance...

Perhaps, but focus is critical and the depth-of-field of the HP lens
that is supplied with the sensor is very narrow (+/-0.2mm).

The laser provides better resolution than an LED, not more range; you
will need to illuminate the target area on your own if you do not use
the supplied optics, too.


Tom

-------------- Original message --------------
From: "sherpadoug" <sherpadoug@...>

Hi, I do a lot of commercial marine robotics and sensors. I am
looking for a better way for a robotic vehicle to track its position
when its wheels are slipping. I am thinking of using the optical
sensor chip from an optical mouse (ADNS-30XX series?) but I need a
couple of inches of ground clearance. Has anyone used these chips
with relay optics so they can track movement at a distance? Are there
any Gotchyas to look out for?

Thanks,
SherpaDoug

Hi, I do a lot of commercial marine robotics and sensors.  I am
looking for a better way for a robotic vehicle to track its position
when its wheels are slipping.  I am thinking of using the optical
sensor chip from an optical mouse (ADNS-30XX series?) but I need a
couple of inches of ground clearance.  Has anyone used these chips
with relay optics so they can track movement at a distance?  Are there
any Gotchyas to look out for?

Thanks,
SherpaDoug

FYI, I just started a forum to help Micromega support their uM-FPU
math coprocessors that I want to use.  Selfishly, I want others to
work with these things, too, so I can pick brains.  I have no
affiliation with the firm.

These are not just floating point calculators.  Programmable.  Matrix
math!  A single instruction FFT!  Might make a great sensor front-end.

http://groups.yahoo.com/group/uMFPU/


Tom

Wow! That's a neat pressure sensor!  Although, it's kinda
unfortunate that there's no breakout board for the I2C version.

May be this article will help you get started with SPI
http://embedded.com/showArticle.jhtml?articleID=9900483
If your have room on your bard, leave pads for pull-up and pull down
resistors for the SDO, SDI and SCLK lines - just in case.


--- In sensorforum@yahoogroups.com, Victor Fraenckel <victorf@...>
wrote:
>
> I am designing a product around the SCP1000 pressure sensor from
VTI
> Technologies:
>
> www.vti.fi
>
> I plan to use the breakout board from Sparkfun:
>
> http://www.sparkfun.com/commerce/product_info.php?products_id=8161
>
> I will be interfacing this device to a ZX-24 Microcontroller from
Elba Corp:
>
> http://www.zbasic.net/
>
> This device is interfaced via the SPI bus which is available on
the
> ZX-24. I have no experience using the SPI bus and would like to
hear
> from any of you that could give me some guidance.
>
> Any enlightenment will be appreciated.
>
> Vic
> --
>
*____________________________________________________________________
________________________*
>
> *Victor Fraenckel
> KC2GUI
> victorf ATSIGN windreader DOT com**
>

I am designing a product around the SCP1000 pressure sensor from VTI
Technologies:

www.vti.fi

I plan to use the breakout board from Sparkfun:

http://www.sparkfun.com/commerce/product_info.php?products_id=8161

I will be interfacing this device to a ZX-24 Microcontroller from Elba Corp:

http://www.zbasic.net/

This device is interfaced via the SPI bus which is available on the
ZX-24. I have no experience using the SPI bus and would like to hear
from any of you that could give me some guidance.

Any enlightenment will be appreciated.

Vic
--

*_______________________________________________________________________________\
_____________*

*Victor Fraenckel
KC2GUI
victorf ATSIGN windreader DOT com**

*

> Maybe each pixel has a pair of LEDs with a common lens.  Like the
2-color LEDs.

Hmmm.  That leads to the question: how well can a green LED detect red
light?

>   Perhaps an optical interrupter across the surface of the control
would work?

That's the path I've been experimenting with recently.  Emmitter on
one side, receiver on the other.  I've casted some small prisms from
urathane resin in order to mount both devices in the panel facing the
operator, and redirect the beam 90 degrees.  It seems promising so
far, although the packaging is still a little 'iffy'.

The required circuit to modulate the signal (to avoid ambient light
interference) and turn the output into a usable format should be
fairly small.  Hopefully, I can use some 0402 or 0603 components, and
cast the whole thing into one piece that I can simply slide over the
control.

Back to the Lab....

-Don

May be, each pixel has a pair of LEDs with a common lens.  Like the
2-color LEDs.

There might be a more brute-force approach to a reflective optical
touch sensor.

An IR LED with lens and IR sensor also with lens (T1 packages, for
example).  They are installed directly under a transparent window in
a front panel so that their optical axes are 90deg with each-other
and 45deg with front panel.  Normally, the brightness patterns of
the LED and the sensor intersect very inefficiently.  But the user's
finger, when present, would reflect some light from the LED to the
sensor and trigger the switch.  To avoid miss-triggering from the
ambient light the LED is driven with a several kHz square wave and
the sensor output is high-pass filtered.

There's still a danger that there will be a stray reflector that
would cause a miss-trigger.  I guess, this can be dealt with some
algorithmic checks.  Additionally, for each "button" there can be
two LED-sensor pairs 90deg (in the front panel plain) with each-
other.

--- In sensorforum@yahoogroups.com, Tom Becker <gtbecker@...> wrote:
>
>
> > http://cs.nyu.edu/~jhan/ledtouch/index.html
>
> Very interesting.  I wonder if the LEDs are switched between
emitting
> and sensing, using neighbors as sources when sensing.  The startup
shows
> a row scan, perhaps self-calibration.
>
>
> Tom
>

Yes, that's how they are doing it.

http://motestruments.com/led-touch-sensor-circuit/

I just threw it out because it's different. I don't know if it would
work in a industrial app. with a guy wearing big black rubber gloves,
but I like the idea.

Chris

> http://cs.nyu.edu/~jhan/ledtouch/index.html

Very interesting.  I wonder if the LEDs are switched between emitting
and sensing, using neighbors as sources when sensing.  The startup shows
a row scan, perhaps self-calibration.


Tom

I ran accross this a couple of days ago. I don't know what it would
take to make a working one. It's an LED touch panel. The video is kink
of big. I have no idea how old the info/link/post is.

http://cs.nyu.edu/~jhan/ledtouch/index.html

HTH, Chris

Thanks for your input Nick,

> By the way, in what area will your device be used?  Industrial?
> Medical?  Marine?

Industrial

> - Thermal sensor.

In this particular application, it won't work, but it's an interesting
idea nonetheless..  I'll have to remember that one.

> - Pneumatic.

hmmm....  I'll have to give that one some thought.  The package
requirements are very strict (read: small), but a pneumatic switch
would allow me to relocate the actual switch mechanism, perhaps to a
location where there's more space available.


> - Hall effect?..

Do Hall Effect sensors sense human contact?  I've never tried it.


Stevech: Great video.  Where can I get one of those?

-Don

Looked up the IP67 on Google and found that it means 2 things:
- Totally protected against dust
- Protected against the effect of immersion between 15cm and 1m
I guess, that means that it's hard to add penetrations to the
enclosure.
By the way, in what area will your device be used?  Industrial?
Medical?  Marine?

Here are some obscure ideas:
- Thermal sensor.  If a user touches the sensor with a warm
finger, it triggers.  However, if the user wears heavy industrial
gloves, it might not work.  Obviously. a thermal sensor will be slow
compared to a mechanical or capacitive.
- Pneumatic.  The user pushes on a small bladder on the
outside of the enclosure.  The bladder communicates through a hose
with a pressure switch inside the enclosure.
- Hall effect?..

Cheers,
Nick

--- In sensorforum@yahoogroups.com, "Don Kirby" <psyclopedia@...>
wrote:
>
> Well, as there seems to be quite a few members registered, and
noone
> is talking, maybe someone would like to chime in on this one.
>
> I've got an existing panel mount control, about 1 square inch,
that I
> would like to add touch sensitivity to.  The space is limited, and
it
> needs to retain its' IP67 rating.
>
> I have tested capacitive sensors, and they work just fine, except
that
> they don't always respond to gloved hands, which the user will most
> likely have.
>
> I could possibly incorporate a resistive type device, but I'm
having
> trouble finding one so small, and I'm not in the position to have
one
> custom manufactured.  Perhaps there is a source for the material in
> bulk that I could cut up myself...
>
> I'm open to ideas, however odd or obscure.
>
> -Don
>

Well, as there seems to be quite a few members registered, and noone
is talking, maybe someone would like to chime in on this one.

I've got an existing panel mount control, about 1 square inch, that I
would like to add touch sensitivity to.  The space is limited, and it
needs to retain its' IP67 rating.

I have tested capacitive sensors, and they work just fine, except that
they don't always respond to gloved hands, which the user will most
likely have.

I could possibly incorporate a resistive type device, but I'm having
trouble finding one so small, and I'm not in the position to have one
custom manufactured.  Perhaps there is a source for the material in
bulk that I could cut up myself...

I'm open to ideas, however odd or obscure.

-Don

GPS is, probably, the sensor that I've studied the most with.  I've
worked for the GPS lab at Stanford.  Right now I work for one of the
GPS startups.  I've also done several GPS-related gigs.

But GPS doesn't work everywhere.  In particular, it doesn't work
underwater...

Nick

I was looking for a sensor-related forum on the web, and I couldn't
find one.  So, I've decided to start this group.  Since sensors are
multidisciplinary objects it usually takes a mix of people to solve
the sensor-related problems.

Cheers,
Nick

Nick, very good idea for a group.

I'm also involved in acoustics, day job and serious hobby.  For fun
I'm doing an acoustical study of snapping shrimp in Monterey Bay CA.

My day job has me designing and fabricating acoustic transducers for
the gas utility industry.

Doug

Nick,
This is a great idea for a discussion group.  Thanks for creating it.
I'm having issues with a GPS receiver, so I'll post my plea for help
soon.
chris