peterhi@syma.sussex.ac.uk (Peter Hickman) (03/20/91)
I'm looking for some references to help me with my second year ( sorry I don't
know what that's called in American ) Artificial Intelligence project. In a
moment of pure insanity I submitted my project as to "Design and build a
system that that shoot clay pigeon" the basis of the design is that everything
can be done with a single camera and no more information than a shooter would
have. There are three broad areas that I have to look at.
1) Looking at the whole sky I need to detect significant notion ( other
than that of the clouds or trees etc ) where in therory a clay pigeon
is flying.
2) Locate the clay in the area of activity and find it's centre.
3) Calculate its trajectory and predict where it will be.
4) Blast the ****** out of the sky.
Point 3 seems to me to be the easiest and point 4 could well be academic as I
have only 15 weeks as of Monday but what I should do for points 1 and 2 is
much more open and I would greatly appreciate any references that you may feel
that could be of help to me. The books I have gone through here on Computer
Vision and Image Processing do not seem to cover the ground that I require in
that they all assume static images of toy worlds. I need a mechanics outlook
on the problems not a theorists at this stage.
Any help greatfully received,
Peter Hi
Peter "You're doing computing as an ARTS degree!" Hickman
COGS U/G PH, University of Sussex, Falmer, Brighton, BN1 9QH
--------------------------------------- peterhi@uk.ac.sussex.syma ------------
"Lego is the Mechano of the intellectual cripple" - Aaron Sloman
--------------------------------------- peterhi@uk.ac.sussex.tsuna -----------
"More beer, more shouting, resistance is useless" - USTA barswilldye@typhoon.unl.edu (03/23/91)
I mailed the following letter to him, but then I thought the rest of you might want to see what I wrote, so here's a copy. ----------------------------------------------------------------------------- To: peterhi@syma.sussex.ac.uk Subject: Re: Shooting pigeons Newsgroups: comp.ai References: <4754@syma.sussex.ac.uk> In comp.ai you write: >In a >moment of pure insanity I submitted my project as to "Design and build a >system that that shoot clay pigeon" the basis of the design is that everything >can be done with a single camera and no more information than a shooter would >have. Start grovelling at the professor (or whatever they call 'em in England) to accept 'signifigant effort' instead of a working system. For YEARS now the military has been developing robotic systems that don't have to do much more than what you're suggesting. It's been said that if anyone developed a system capable of playing ping-pong (table tennis), it would be locked up and classified. You have two major problems. One is a computer vision problem of identifying and tracking the target IN REAL TIME. This is non-trivial, even when you have artificial constraints put in to make sure the signal-to-noise ratio is high. The second problem is building the mechanics necessary to move the shotgun ACCURATELY and QUICKLY to the desired position. This is also non-trivial, and also quite a different problem from the software-intensive vision problem. I'm not saying don't do it. In fact, I'd LOVE to work on something like this. I'm just saying start small and don't have illusions about the task before you. I suggest starting with a simplified software simulation. Have one computer (or a camera) send a video picture of a black ball moving in a straight line against a white background. See if you can keep crosshairs on the ball. Then use a ball with a non-straight flight path, and try to LEAD the target in real time. It may help to look up some basic references to radar tracking algorithms, stuff like aquisition gates (large area around the target), tracking gate (small area around target), and turn gates (in-between sized gate). Don't try a fancy system with a high signal-to-noise ratio until you've got the simple system working. The S/N ratio problem can never truly be 'solved', we just get better and better at it. Good luck. Please tell me how it worked out. William L. Dye willdye@typhoon.unl.edu
loren@ingrid.llnl.gov (Loren Petrich) (03/23/91)
In article <4754@syma.sussex.ac.uk> peterhi@syma.sussex.ac.uk (Peter Hickman) writes:
:I'm looking for some references to help me with my second year ( sorry I don't
:know what that's called in American ) Artificial Intelligence project. In a
:moment of pure insanity I submitted my project as to "Design and build a
:system that that shoot clay pigeon" the basis of the design is that everything
:can be done with a single camera and no more information than a shooter would
:have. There are three broad areas that I have to look at.
You might want to check out alternatives to traditional AI,
such as Neural Nets, Fuzzy Logic, and the kind of robotics that Brooks
has used in his mechanical insects. I make these plugs because these
are techniques that have produced _results_, and that's what one is
supposed to get, right?
: 1) Looking at the whole sky I need to detect significant notion ( other
: than that of the clouds or trees etc ) where in therory a clay pigeon
: is flying.
Interesting question in Artificial Vision.
To look for a Clay Pigeon, one need only look for something
that moves at a different speed from the background. This is easy to
detect for a static system, but that cannot be taken for granted. One
needs to compare successive images, and find out what has changed
abnormally.
Frogs are known to have "bug detector" neurons in their brains
which fire whenever a small object moves past. [if I remember
correctly] A computerized "bug detector" might be a valuable
innovation in artificial vision.
Detecting the motion of small objects may be a similar challenge.
: 2) Locate the clay in the area of activity and find it's centre.
Part of above.
: 3) Calculate its trajectory and predict where it will be.
One might want some strategy that learns from experience; a
Neural Net strategy, let us say. One minimizes the error function of
miss distance; one's inputs are the Clay Pigeon's position and
velocity and the positioning of one's gun.
One can train by shooting several times and noticing the miss
distances. One trains to minimize the function (CP position, CP
velocity) -> (miss distance). Zero miss distance is, of course, what
is desired.
: 4) Blast the ****** out of the sky.
Easy if one has succeeded in training for (3).
:Point 3 seems to me to be the easiest and point 4 could well be academic as I
:have only 15 weeks as of Monday but what I should do for points 1 and 2 is
:much more open and I would greatly appreciate any references that you may feel
:that could be of help to me. The books I have gone through here on Computer
:Vision and Image Processing do not seem to cover the ground that I require in
:that they all assume static images of toy worlds. I need a mechanics outlook
:on the problems not a theorists at this stage.
One may want to check out some of the papers in the IJCNN
Conferences on Neural Networks(?).
Sorry, I don't know of any elementary NN textbooks.
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
Loren Petrich, the Master Blaster: loren@sunlight.llnl.gov
Since this nodename is not widely known, you may have to try:
loren%sunlight.llnl.gov@star.stanford.eduhm02+@andrew.cmu.edu (Hans P. Moravec) (03/23/91)
> loren@ingrid.llnl.gov (Loren Petrich) writes: > You might want to check out alternatives to traditional AI, > such as Neural Nets, Fuzzy Logic, and the kind of robotics that Brooks > has used in his mechanical insects. I make these plugs because these > are techniques that have produced _results_, and that's what one is > supposed to get, right? Expert systems are not used much in robot vision systems, but "traditional" methods such as high pass, low pass and specialized filters, correlation, frequency transforms, and specialized operators are in practical use in industrial robots and such weapons as the Tomahawk cruise missile and fire and forget TV guided bombs and missles, where millions of pixels of image data are processed per second. The neo-cybernetic techniques have yet to prove themselves in such computationally difficult domains. Fuzzy logic, neural nets and reactive organizations have so far been used only in relatively simple contexts, like camera autofocus and train speed control, and a few mobile robots, where the input is limited to a few dozen measurements at a time. -- Hans Moravec
karln@uunet.uu.net (03/26/91)
In article <4754@syma.sussex.ac.uk> peterhi@syma.sussex.ac.uk (Peter Hickman) writes: >moment of pure insanity I submitted my project as to "Design and build a >system that that shoot clay pigeon" the basis of the design is that everything >can be done with a single camera and no more information than a shooter would >have. There are three broad areas that I have to look at. I did not know this was the rec.humor conference :) > > 1) Looking at the whole sky I need to detect significant notion ( other > than that of the clouds or trees etc ) where in therory a clay pigeon > is flying. Do you mean that if the program gets a significant 'notion' it'll open fire? Weeee, I wish i was a student again ... > > 2) Locate the clay in the area of activity and find it's centre. > So you'll have to open up with your twelve-gauge on anything that is about 3 inches long by 1 inch high and moving .... > > 3) Calculate its trajectory and predict where it will be. And then shoot at it with devastating accuracy. > > 4) Blast the ****** out of the sky. > I *really* hope this is a video game. At least a simulation? Salt maybe? > >Point 3 seems to me to be the easiest and point 4 could well be academic as I >have only 15 weeks as of Monday but what I should do for points 1 and 2 is >much more open and I would greatly appreciate any references that you may feel >that could be of help to me. The books I have gone through here on Computer >Vision and Image Processing do not seem to cover the ground that I require in >that they all assume static images of toy worlds. I need a mechanics outlook >on the problems not a theorists at this stage. > >Any help greatfully received, > >Peter Hi ^^-----> Very ... Peter, I do wish you much luck with this project. Can I take out a life insurance policy on you? I hope your device is familiar with Asimovs rules for robots 1: They shall not though action or in-action allow a human come to harm. 2: They shall not though action or inaction allow a human to harm another. 3: They shall not though action or inaction allow themselves harm. (i think). Karl Nicholas karln!karln@uunet.uu.net
ta2cs220@troi.cc.rochester.edu (Hossein Firooznia) (03/26/91)
In <1991Mar25.171303.17211@uunet.uu.net> karln@uunet.uu.net writes: > Peter, I do wish you much luck with this project. Can I take out a life >insurance policy on you? I hope your device is familiar with Asimovs rules >for robots >1: They shall not though action or in-action allow a human come to harm. >2: They shall not though action or inaction allow a human to harm another. >3: They shall not though action or inaction allow themselves harm. >(i think). I think the second law is: They shall obey commands given by humans, unless this violates the first law. Each law has priority over those following it. -Hoss H. Y. Firooznia hfir_ltd@uhura.cc.rochester.edu
ahlenius@motcid.UUCP (Mark Ahlenius) (03/26/91)
loren@ingrid.llnl.gov (Loren Petrich) writes: some stuff deleted... > You might want to check out alternatives to traditional AI, >such as Neural Nets, Fuzzy Logic, and the kind of robotics that Brooks >has used in his mechanical insects. I make these plugs because these >are techniques that have produced _results_, and that's what one is >supposed to get, right? more stuff deleted... > To look for a Clay Pigeon, one need only look for something >that moves at a different speed from the background. This is easy to even more stuff deleted... All right, I for one am tired of this decoy constantly bashing AI in this group. If you are that opposed to what you consider AI techniques, why don't you spend your efforts in another group? Granted there were people in the AI arena that said that this technology would produce "human-like" machines but there are fanatics in every field that exists. I for one do not really believe that we shall see such machines come to existence, but then again we do let our imaginations run wild sometimes. So if you look at the big picture, yes for some 30 odd years we do not have these human-like machines - but we do have many other facets gleaned from the technology. Per previous net conversations many of the advances that "AI" techniques have brought to the party have been discussed. Many of these are in constant use today. And many neural net and fuzzy logic applications are being applied today as well. These are small individual techniques that each solve small portions of big problems. But I am perfectly satisfied applying such techniques to computer systems to help them act in a more intelligent fashion. And I will apply different techniques be they neural nets, fuzzy, expert systems, heuristic search, where ever the technology fits the problem best. -- =============== regards 'mark ============================================= Mark Ahlenius voice:(708)-632-5346 email: uunet!motcid!ahleniusm Motorola Inc. fax: (708)-632-2413 Arlington, Hts. IL, USA 60004
moskowit@paul.rutgers.edu (Len Moskowitz) (03/26/91)
I'm surprised that no one has pointed out that this problem has been addressed and solved very nicely using traditional ( non-AI) technologies. No AI is needed to track a ballistic target, determine its centroid, and to aim and fire a weapon. Why re-invent the wheel? Len Moskowitz
ted@nmsu.edu (Ted Dunning) (03/27/91)
In article <Mar.26.09.51.32.1991.18229@paul.rutgers.edu> moskowit@paul.rutgers.edu (Len Moskowitz) writes:
I'm surprised that no one has pointed out that this problem has been
addressed and solved very nicely using traditional ( non-AI)
technologies. No AI is needed to track a ballistic target, determine
its centroid, and to aim and fire a weapon.
well, I'm surprised that nobody has pointed out that for shooting clay
pigeons, you don't even have to track them, you just have to point the
gun where they are _going_ to be, and pull the trigger at the right
time.
if you are willing to take an `engineering' approach to the problem as
opposed to an `ai' approach, then you can set up a simple
photoelectric detector which will determine when the pigeon is
released, wait for a small bit, and pull the trigger.
and, of course, this is cheating. that is what engineering is all
about, redefining impossible problems so they have practical
solutions. loren@ingrid.llnl.gov (Loren Petrich) (03/27/91)
In article <TED.91Mar26123325@ithaka.nmsu.edu> ted@nmsu.edu (Ted Dunning) writes: >In article <Mar.26.09.51.32.1991.18229@paul.rutgers.edu> moskowit@paul.rutgers.edu (Len Moskowitz) writes: > I'm surprised that no one has pointed out that this problem has been > addressed and solved very nicely using traditional ( non-AI) > technologies. No AI is needed to track a ballistic target, determine > its centroid, and to aim and fire a weapon. Tracking and aiming may be relatively straightforward, but one has to first recognize the target. And that's where the AI comes in. >well, I'm surprised that nobody has pointed out that for shooting clay >pigeons, you don't even have to track them, you just have to point the >gun where they are _going_ to be, and pull the trigger at the right >time. But that is EXACTLY what one has to find out in the first place. Where the Clay Pigeon is. >if you are willing to take an `engineering' approach to the problem as >opposed to an `ai' approach, then you can set up a simple >photoelectric detector which will determine when the pigeon is >released, wait for a small bit, and pull the trigger. >and, of course, this is cheating. that is what engineering is all >about, redefining impossible problems so they have practical >solutions. But what if one could not set up a "clay pigeon detector" in this fashion? That's what the original posting was all about. And cheating is cheating. $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ Loren Petrich, the Master Blaster: loren@sunlight.llnl.gov Since this nodename is not widely known, you may have to try: loren%sunlight.llnl.gov@star.stanford.edu
demers@odin.ucsd.edu (David E Demers) (03/28/91)
In article <19321@lanl.gov> varga@vxcrna writes: > >In <willdye.669661816@typhoon> willdye@typhoon.unl.edu writes: > >> It's been said that if anyone developed a system >> capable of playing ping-pong (table tennis), it would be locked up >> and classified. > >Five years ago I visited the University of Tampere in Finland and saw a >machine built to play ping-pong. They should be contacted and asked for >their experience. See Russell Andersson's book, A Robot Ping-Pong Player: Experiment in Real-Time Intelligent Control (MIT Press, 1988). It's not locked up and classified. It is of course not capable of playing at human expert level, but it DOES play robot ping-pong. Dave -- Dave DeMers demers@cs.ucsd.edu Computer Science & Engineering C-014 demers%cs@ucsd.bitnet UC San Diego ...!ucsd!cs!demers La Jolla, CA 92093-0114 (619) 534-8187,-0688 ddemers@UCSD
fnwlr1@acad3.alaska.edu (RUTHERFORD WALTER L) (03/28/91)
In article <Mar.26.09.51.32.1991.18229@paul.rutgers.edu>, moskowit@paul.rutgers.edu (Len Moskowitz) writes... >I'm surprised that no one has pointed out that this problem has been >addressed and solved very nicely using traditional ( non-AI) >technologies. No AI is needed to track a ballistic target, determine >its centroid, and to aim and fire a weapon. > I was also waiting for someone else to point out that we already have a system to do this and more... they call it Patroit. Good luck getting the designers to answer technical questions though. :-) --------------------------------------------------------------------- Walter Rutherford P.O. Box 83273 \ / Computers are NOT intelligent; Fairbanks, Alaska 99708 - X - / \ they just think they are! fnwlr1@acad3.alaska.edu ---------------------------------------------------------------------
varga@vxcrna (VARGA,LASZ./PS) (03/28/91)
In <willdye.669661816@typhoon> willdye@typhoon.unl.edu writes: > It's been said that if anyone developed a system > capable of playing ping-pong (table tennis), it would be locked up > and classified. Five years ago I visited the University of Tampere in Finland and saw a machine built to play ping-pong. They should be contacted and asked for their experience. Laszlo Zsolt Varga ======================================================================= Letters: CERN, CH-1211 GENEVA 23, SWITZERLAND ; Phone: +41 22 767 5971 Fax: +41 22 785 0515 ; Tx: 419000 CER CH ; E-mail: varga@cernvm.cern.ch =======================================================================
schraag@fwi.uva.nl (Joost Schraag) (03/28/91)
In article <1991Mar27.215225.25491@ims.alaska.edu> (of comp.ai)
fnwlr1@acad3.alaska.edu wrote:
+ In article <Mar.26.09.51.32.1991.18229@paul.rutgers.edu>, moskowit@paul.rutgers.edu (Len Moskowitz) writes...
+ >I'm surprised that no one has pointed out that this problem has been
+ >addressed and solved very nicely using traditional ( non-AI)
+ >technologies. No AI is needed to track a ballistic target, determine
+ >its centroid, and to aim and fire a weapon.
+ >
+
+ I was also waiting for someone else to point out that we already have
+ a system to do this and more... they call it Patroit. Good luck getting
+ the designers to answer technical questions though. :-)
+
+
+ ---------------------------------------------------------------------
+ Walter Rutherford
+ P.O. Box 83273 \ / Computers are NOT intelligent;
+ Fairbanks, Alaska 99708 - X -
+ / \ they just think they are!
+ fnwlr1@acad3.alaska.edu
+ ---------------------------------------------------------------------
Another system is the GoalKeeper, made by Holland Signaal. It is a high
speed Gatling Gun, and is able to intercept Exocet-rockets. Dutch navy-
ships use them when the Exocet goes through the first anti-missile-missile
defence. The GoalKeeper shoots 4200 grenates a minute in the path of the
Exocet, causing it to fly into a wall of steel.
They do use a realtime tracking system, however i think the details are
secret. Sorry, i don't have the adress of Holland Signaal.
Using this gun on clay-pigeons will have a high strike-rate.
joost.
--
Joost Schraag e-mail: schraag@fwi.uva.nl ICBM: 52.22`N 04.54`EG.Joly@cs.ucl.ac.uk (Gordon Joly) (03/31/91)
Len Moskowitz writes: > I'm surprised that no one has pointed out that this problem has been > addressed and solved very nicely using traditional ( non-AI) > technologies. No AI is needed to track a ballistic target, determine > its centroid, and to aim and fire a weapon. > > Why re-invent the wheel? > > > Len Moskowitz Yeah; for sure. I never did like the notion that robotics was part of AI;-) Gordon Joly +44 71 387 7050 ext 3716 Internet: G.Joly@cs.ucl.ac.uk UUCP: ...!{uunet,ukc}!ucl-cs!G.Joly Computer Science, University College London, Gower Street, LONDON WC1E 6BT "I didn't do it. Nobody saw me do it. You can't prove anything!"
G.Joly@cs.ucl.ac.uk (Gordon Joly) (04/01/91)
And to add to that all that, I watch my dog get the ball from the letter box (about two to three times it own height) and the birds land on the (often spinning) peanut feeder. These are neat tricks that will take some duplication/simulation. Gordon.
cam@aifh.ed.ac.uk (Chris Malcolm) (04/10/91)
In article <1524@ucl-cs.uucp> G.Joly@cs.ucl.ac.uk (Gordon Joly) writes: >And to add to that all that, I watch my dog get the ball from the >letter box (about two to three times it own height) and the birds land >on the (often spinning) peanut feeder. >These are neat tricks that will take some duplication/simulation. Ever watch squirrels jumping from tree to tree? They run out along one branch, right out to the twiggy end, which is bending under their weight, and without seeming to pause in their run, make a JUMP which leaves the exit branch springing, and make it no bother to the target branch: they compensate for the spring in the branch they're leaving. But nothing beats gibbons flying through the trees with nonchalant acrobatics of incredible grace and impossible skill! -- Chris Malcolm cam@uk.ac.ed.aipna +44 (0)31 667 1011 x2550 Department of Artificial Intelligence, Edinburgh University 5 Forrest Hill, Edinburgh, EH1 2QL, UK DoD #205