Newsgroups: comp.robotics
Path: brunix!cat.cis.Brown.EDU!agate!howland.reston.ans.net!vixen.cso.uiuc.edu!sdd.hp.com!hplabs!hplntx!hparc0.aus.hp.com!robink
From: robink@aus.hp.com (Robin Kenny)
Subject: Re: PROPOSAL: A robotic Butterlfy!
Sender: news@hparc0.aus.hp.com (News Adminstrator)
Message-ID: <1994Aug10.023633.3404@hparc0.aus.hp.com>
Date: Wed, 10 Aug 1994 02:36:33 GMT
References: <1994Jul28.153808.22030@msuvx2.memphis.edu> <ELEQIM.15.0009A00F@herts.ac.uk> <1994Jul29.103802.22064@msuvx2.memphis.edu> <ELEQIM.16.000A3A36@herts.ac.uk>
Nntp-Posting-Host: hatchet.aus.hp.com
Organization: HP Australasian Response Centre (Melbourne)
X-Newsreader: TIN [version 1.2 PL2]
Lines: 49

My apologies for missing the original post, however here are a couple of
observations and a suggestion or two...

Ian Munro (ELEQIM@herts.ac.uk) wrote:
: >Perhaps you (and several others) missed the point.  All I want the thing to
: >do is fly STRAIGHT UP when light is applied, and DOWN as the light source
: >decreases. 
  This surprises me as you are describing a negative feedback loop - light
intensity decreases with distance, so if the light is above the robot we
are talking run-away upwards or downwards flight, depending on whatever
random fluctuation starts the feedback. (Of course the original post may
have mentioned a side or underneath mounted light, sorry)

Self powered flight straight up is indeed difficult; but what of a funnel
of air being forced directly up under the robot and then all it has to do
is "coast" or "tack" the airstream like a boat does? It may even be possible 
for the robot to wait for a suitable random air current moving where it wants 
to go before flying.

For two-way telemetry I have often thought that a SUPER LOW POWER link 
could be made by using a transmissive LCD shutter over the top of some
retroreflective layer, like 3M Retro Tape, on the robot body. 
"Retroreflective" means that a light ray will reflect EXACTLY back at the
light source. This is the stuff used on those road signs that really 
light up at night. Animal eyes do this too. You need to have your viewing 
eye right next to a flashlight when you try it. 

The LCD can be blinked on and off (dark and clear) with a slow serial stream.
Most LCDs can be strobed to about 7 Hz without consumming too much power. 
The light source needs a hooded sensor near on-axis to the outgoing
beam and aimed at the robot to see the reflected light being blinked 
on and off like a semaphore.
(Late thought - why not modulate the outgoing light, like IR remote controls
 do? Then a 40kHz return modulated at about 7Hz could be seen by a wider-angle
 IR sensor ($5 at Radio Shack!) and the requirement for tracking the robot
 may be not necessary)

A full scale pterodactyl was built about two years ago to prove that
pterodactyls could indeed fly, and that they could do so with the muscles
and bones paleontologists calculated they had. It had to feature realistic
skin because at the time a discovery was made that some species actually
had fur! I forget the conclusions, but a quick summary was that they 
were more gliders than power flyers. This was fine as the nest sites were
often on windy cliff faces.

Robin Kenny - robink@hparc0.aus.hp.com           "New ideas in status quo"
(everything in this message is PERSONAL OPINION ONLY and has no connection
with my work or my employer, the Hewlett-Packard Company Australia)
--------------------------------------------------------------------------
