From ???@??? Sun Nov 19 12:51:45 1995 Received: from endor.harvard.edu by rz1.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA10072; Mon, 13 Nov 1995 12:25:01 +0100 Received: from netop3.harvard.edu (netop3.harvard.edu [128.103.205.103]) by endor (8.6.12/8.6.12) with SMTP id GAA18454 for ; Mon, 13 Nov 1995 06:05:45 -0500 Received: by netop3.harvard.edu (5.65/DEC-Ultrix/4.3) id AA18237; Mon, 13 Nov 1995 06:00:08 -0500 Received: from KITES-NEWS-GATEWAY by netop3.harvard.edu with netnews for kites-news@das.harvard.edu (kites@das.harvard.edu) To: kites@das.harvard.edu Date: Mon, 13 Nov 1995 08:41:16 +0100 From: buchholz@fbm.hs-bremen.de (Jvrg J. Buchholz) Message-Id: <9511130741.AA11137@rz1.rz.hs-bremen.de> Organization: Division of Applied Sciences, Harvard University Sender: kites-request@das.harvard.edu Subject: Autopilot for Revolution (AutoRev) X-Mozilla-Status: 0001 Content-Length: 2785 Dear friends Five weeks ago they gave me this job as a professor for flight control at the Hochschule Bremen, Germany. One of the advantages of such a business is the freedom to choose your own research topic: I will invest my scientific energy into the modelling, simulation and control of kites. Some of you might know that about two years ago I gave birth to a real-time pilot-in-the-loop simulator of a Revolution kite by the name of RevSim. My next two years project will now be a Revolution autopilot (AutoRev), i.e. a box in the grass controlling all four lines of a rev. It will use a video camera and pattern recognition algorithms to receive information about the state (attitude and position) of the kite and utilize the mathematical model of RevSim for a model following control algorithm in the control computer. This will enable me to predefine a complete ballet in the computer (e.g. via Kiteflight) in advance and then let the autopilot win championships against human competitors. Isn't that a challenge? Will AutoRev ever do the Axel? I am totally aware of the fact that no kite flyer needs a kite flying machine. _We_ want to have fun with our kites, _we_ want to fly them ... So, don't take the paragraph above too seriously. But - I believe that such a project could be a perfect opportunity to promote the colorful idea of kiting in those gray academic areas of science and research. Last week I gave a lecture on RevSim and the mathematical model behind it at an international symposium on real-time simulation in Delft. You should have seen those scientists and managers, who usually talk about A320 and space simulators and investments of millions of dollars, suddenly becoming interested in the fascinating world of rev flying (I showed part of the Revolution teaching video). To make the long story short: Right now I am looking for __serious__ applications of such an AutoRev. I still have to find partners in industry and ministry to support these research activities. The Hochschule Bremen is not willing to fund some $30.000 just for the fun of it. I was thinking for example of some environmental applications to use an autonomously flying kite based on a ship for the detection of maritime waste (oil, ...). Could you help me out with other AutoRev applications that might persuade the right people ... Thank you very much -- buchholz@fbm.hs-bremen.de ****************************************************************** This message was written by the kites/rec.kites mail/news gateway. To post messages to rec.kites, send mail to: "kites@das.harvard.edu" To make list requests and inquiries send mail to "kites-request@das.harvard.edu". From ???@??? Sun Nov 19 12:51:51 1995 Received: from endor.harvard.edu by rz1.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA09700; Tue, 14 Nov 1995 10:45:23 +0100 Received: from netop3.harvard.edu (netop3.harvard.edu [128.103.205.103]) by endor (8.6.12/8.6.12) with SMTP id MAA24612 for ; Mon, 13 Nov 1995 12:09:47 -0500 Received: by netop3.harvard.edu (5.65/DEC-Ultrix/4.3) id AA00708; Mon, 13 Nov 1995 12:00:32 -0500 Received: from KITES-NEWS-GATEWAY by netop3.harvard.edu with netnews for kites-news@das.harvard.edu (kites@das.harvard.edu) To: kites@das.harvard.edu Date: 13 Nov 1995 11:05:50 GMT From: Gregor Westemeier Message-Id: <4878qe$6m@spinne.ip.lu> Organization: Infopartners Sender: kites-request@das.harvard.edu References: <9511130741.AA11137@rz1.rz.hs-bremen.de> Subject: Re: Autopilot for Revolution (AutoRev) X-Mozilla-Status: 0011 Content-Length: 583 always think big brother (no, not the kite). what a about an autonomous video control system for hazardous terrains? you would need to make the kite bigger to lift the stuff, but this is rather an advantage, as the kite becomes slower *what the h[e,a]ck? it's -grg-* ****************************************************************** This message was written by the kites/rec.kites mail/news gateway. To post messages to rec.kites, send mail to: "kites@das.harvard.edu" To make list requests and inquiries send mail to "kites-request@das.harvard.edu". From ???@??? Sun Nov 19 12:51:52 1995 Received: from endor.harvard.edu by rz1.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA07470; Tue, 14 Nov 1995 12:22:09 +0100 Received: from netop3.harvard.edu (netop3.harvard.edu [128.103.205.103]) by endor (8.6.12/8.6.12) with SMTP id MAA24652 for ; Mon, 13 Nov 1995 12:10:41 -0500 Received: by netop3.harvard.edu (5.65/DEC-Ultrix/4.3) id AA00651; Mon, 13 Nov 1995 12:00:12 -0500 Received: from KITES-NEWS-GATEWAY by netop3.harvard.edu with netnews for kites-news@das.harvard.edu (kites@das.harvard.edu) To: kites@das.harvard.edu Date: 13 Nov 1995 14:09:21 GMT From: salanne@convex.csc.FI (Simo Salanne) Message-Id: <487jih$9vk@ankka.csc.fi> Organization: STACK Finland Sender: kites-request@das.harvard.edu References: <9511130741.AA11137@rz1.rz.hs-bremen.de> Subject: Re: Autopilot for Revolution (AutoRev) X-Mozilla-Status: 0011 Content-Length: 1458 In <9511130741.AA11137@rz1.rz.hs-bremen.de> buchholz@fbm.hs-bremen.de (Jvrg J. Buchholz) writes: >Dear friends >My next two years project will now be a Revolution autopilot (AutoRev), >i.e. a box in the grass controlling all four lines of a rev. >It will use a video camera and pattern recognition algorithms to receive >information about the state (attitude and position) of the kite and >utilize the mathematical model of RevSim for a model following control >algorithm in the control computer. >This will enable me to predefine a complete ballet in the computer >(e.g. via Kiteflight) in advance and then let the autopilot win >championships against human competitors. In STACK Rules and Regulations Committee we have thought that one have to be a "licensed competitor" to be able to compete. Apparently we need to add a statement, that robots (I don't mean Robos :-) don't qualify as licensed competitors;-/ Simo -- Simo.Salanne@csc.fi STACK Finland & STARRC Chair ==================================================================== http://www.kfs.org/kites/simo/simo.html ****************************************************************** This message was written by the kites/rec.kites mail/news gateway. To post messages to rec.kites, send mail to: "kites@das.harvard.edu" To make list requests and inquiries send mail to "kites-request@das.harvard.edu". From ???@??? Sun Nov 19 12:51:50 1995 Received: from endor.harvard.edu by rz1.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA09422; Tue, 14 Nov 1995 10:34:09 +0100 Received: from netop3.harvard.edu (netop3.harvard.edu [128.103.205.103]) by endor (8.6.12/8.6.12) with SMTP id SAA05674 for ; Mon, 13 Nov 1995 18:07:52 -0500 Received: by netop3.harvard.edu (5.65/DEC-Ultrix/4.3) id AA15910; Mon, 13 Nov 1995 18:00:07 -0500 Received: from KITES-NEWS-GATEWAY by netop3.harvard.edu with netnews for kites-news@das.harvard.edu (kites@das.harvard.edu) To: kites@das.harvard.edu Date: 13 Nov 1995 15:45:08 GMT From: Rodger Duffett Message-Id: <487p64$ecp@ucthpx.uct.ac.za> Organization: UCT Sender: kites-request@das.harvard.edu References: <9511130741.AA11137@rz1.rz.hs-bremen.de> Subject: Re: Autopilot for Revolution (AutoRev) X-Mozilla-Status: 0011 Content-Length: 906 Nice Job! I have long dreamed of a computerised autopilot to control a power-kite on a boat. Only this would most likely not be a Rev type kite. I think that it would be something like a super efficient, lightweight but very strong glider tethered to the boat by a single 'string'! Cheerio -- __________________________________________________________________ Rodger Duffett Internet Address: rodger@ray.uct.ac.za Dept. Radiation Oncology Telephone : +2721 404 3135 Radiobiology Section University of Cape Town Groote Schuur Hospital South Africa ****************************************************************** This message was written by the kites/rec.kites mail/news gateway. To post messages to rec.kites, send mail to: "kites@das.harvard.edu" To make list requests and inquiries send mail to "kites-request@das.harvard.edu". From ???@??? Sun Nov 19 12:51:46 1995 Received: from ns.dknet.dk by rz1.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA09353; Tue, 14 Nov 1995 09:27:52 +0100 Received: from login.dknet.dk (root@login.dknet.dk [193.88.44.43]) by ns.dknet.dk (8.6.12/8.6.12) with SMTP id RAA05385 for ; Mon, 13 Nov 1995 17:03:51 +0100 Received: by login.dknet.dk (4.1/SMI-4.1DKnet00) id AA09219; Mon, 13 Nov 95 16:59:40 +0100 From: rolf@login.dknet.dk (Rolf V. Oestergaard) To: buchholz@fbm.hs-bremen.de.v.v (J\vrg J. Buchholz) Subject: Re: Autopilot for Revolution (AutoRev) Newsgroups: rec.kites Date: Mon, 13 Nov 1995 16:47:37 -0100 Organization: DKnet / EUnet Denmark - Login Tjenesten Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Message-Id: <5S4pwMbaARtB088yn@login.dknet.dk> References: <9511130741.AA11137@rz1.rz.hs-bremen.de> In-Reply-To: <9511130741.AA11137@rz1.rz.hs-bremen.de> Lines: 41 X-Mozilla-Status: 0011 Content-Length: 1924 On rec.kites, You wrote: >Five weeks ago they gave me this job as a professor for flight control >at the Hochschule Bremen, Germany. One of the advantages of such a >business is the freedom to choose your own research topic: I will >invest my scientific energy into the modelling, simulation and >control of kites. Congratulations. >My next two years project will now be a Revolution autopilot (AutoRev), >i.e. a box in the grass controlling all four lines of a rev. >It will use a video camera and pattern recognition algorithms to receive >information about the state (attitude and position) of the kite and >utilize the mathematical model of RevSim for a model following control >algorithm in the control computer. Sounds very interesting. Please keep us informed, as You progress through the project. I myself have been thinking on and off about a similar project, but never had the time for the complete dedication, that is really needed. If You would like to have a different view of some wierd idea You cook up, feel free to send it for comments. >To make the long story short: Right now I am looking for __serious__ >applications of such an AutoRev. I still have to find partners in >industry and ministry to support these research activities. >The Hochschule Bremen is not willing to fund some $30.000 just for >the fun of it. The only similar thing I have heard of was, was a R/C chopper to carry a test antenna to perform a complete measurement of a new antenna installation. Plotting the 3D field strength of the antenna. The chopper was to be controlled by a computer servo system. If I come to think of something You could use, I will let You know. See You, ______________ Rolf V. Oestergaard, Copenhagen, Denmark, M.Sc.EE \ /\ / email: rolf@login.dknet.dk \/ \/ From ???@??? Sun Nov 19 12:51:48 1995 Received: from endor.harvard.edu by rz1.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA09400; Tue, 14 Nov 1995 10:30:09 +0100 Received: from netop3.harvard.edu (netop3.harvard.edu [128.103.205.103]) by endor (8.6.12/8.6.12) with SMTP id AAA11079 for ; Tue, 14 Nov 1995 00:10:33 -0500 Received: by netop3.harvard.edu (5.65/DEC-Ultrix/4.3) id AA22257; Tue, 14 Nov 1995 00:00:20 -0500 Received: from KITES-NEWS-GATEWAY by netop3.harvard.edu with netnews for kites-news@das.harvard.edu (kites@das.harvard.edu) To: kites@das.harvard.edu Date: 13 Nov 1995 19:31:31 GMT From: daveculp@bdt.com (Dave Culp) Message-Id: Organization: Beckemeyer Development, Oakland CA Sender: kites-request@das.harvard.edu References: <9511130741.AA11137@rz1.rz.hs-bremen.de> Subject: Re: Autopilot for Revolution (AutoRev) X-Mozilla-Status: 0011 Content-Length: 3832 In article <9511130741.AA11137@rz1.rz.hs-bremen.de>, buchholz@fbm.hs-bremen.de (Jvrg J. Buchholz) wrote: It will use a video camera and pattern recognition algorithms to receive > information about the state (attitude and position) of the kite and > utilize the mathematical model of RevSim for a model following control > algorithm in the control computer. Have you considered placing reflectors (bicycle-type) on each wing tip and using a scanning laser (supermarket-type) to map the wing tips in real time? Should be much easier on your processor. If you color or bar coded the reflectors, you could control multiple kites simultaneously. > > This will enable me to predefine a complete ballet in the computer > (e.g. via Kiteflight) in advance and then let the autopilot win > championships against human competitors. Isn't that a challenge? Of course it's not a challenge! You could trail a ribbon tail and write your name in the sky with it! (Hey, how about attaching smoke geneator and going sky writing? > To make the long story short: Right now I am looking for __serious__ > applications of such an AutoRev. I still have to find partners in > industry and ministry to support these research activities. > The Hochschule Bremen is not willing to fund some $30.000 just for > the fun of it. The above scheme was postulated 10-15 years ago as a way of controlling (very) large kites for ship propulsion (!) The fly in the ointment was that NATO countries make it illegal (or did then) to aim a laser above the horizon--could blind an aircraft pilot. I think political climates; certainly laser technology, have changed since then--should be possible now. Possible kite uses: * Bird scarer; kite just "window wipes" back and forth accross field. Random motion in flight path keeps the pattern novel and effective. * Hoisting ultra low volume crop sprayer nozzle behind tractor for poor country's version of aerial spraying. * Wartime camera (or other electronics) platform. Imagine putting Rev on "drunkard's walk" flight path, then trying to hit it with anything. * Fish finder. Tuna trawlers often mount small helicopters to get high enough to spot schools of fish. Camera on a kite makes it simple. Trawler makes wind. * Ditto (as you suggested) for spotting dirty offshore dumping. * Spin-off: If you get your grant and develop the beastie, I can certainly use it to control Flexifoils on my speed sailboats. Imagine what one could do with a *really* controllable, fast kite (I do!) * Of course, control of large kite for ship/small craft propulsion. Years ago, it was postulated that a kite powered submarine chaser could be built with absolutely no metal, and no propulsion noises. Ditto for a torpedo delivery system (kinda gruesome, but military's where the $$s are) * I wonder if the technology would carry over to parachute controls. Gives a new meaning to "Airshow" maybe? I suspect that entertainment value, (like "laser light shows" did with controllable lasers) would be most popular usage. Imagine a computer controlled 6-kite ballet, with 6" separation on maneuvers... -- ***************************************************************** Dave Culp daveculp@bdt.com Diablo Realty http://www.bdt.com/home/daveculp/ 975 Ygnacio Valley Road 510.933.9300 Walnut Creek, CA 94596 510.933.9305 x 308 ***************************************************************** ****************************************************************** This message was written by the kites/rec.kites mail/news gateway. To post messages to rec.kites, send mail to: "kites@das.harvard.edu" To make list requests and inquiries send mail to "kites-request@das.harvard.edu". From ???@??? Sun Nov 19 12:51:46 1995 Received: from ngeout01.news.aol.com by rz1.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA07518; Tue, 14 Nov 1995 08:32:43 +0100 Received: from newsbf02.news.aol.com (newsbf02.mail.aol.com [152.163.172.50]) by ngeout01.news.aol.com (8.6.12/8.6.12) with SMTP id XAA01018 for ; Mon, 13 Nov 1995 23:18:03 -0500 From: kitestuff@aol.com Received: by newsbf02.news.aol.com (1.38.193.5/16.2) id AA28505; Mon, 13 Nov 1995 23:17:07 -0500 Date: Mon, 13 Nov 1995 23:17:07 -0500 Posted-Date: Mon, 13 Nov 1995 23:17:07 -0500 Received-Date: Mon, 13 Nov 1995 23:17:07 -0500 Message-Id: <9511140417.AA28505@newsbf02.news.aol.com> Reply-To: kitestuff@aol.com (KITESTUFF) To: buchholz@fbm.hs-bremen.de.v (J\vrg J. Buchholz) Subject: Re: Autopilot for Revolution (AutoRev) X-Organization: America Online, Inc. (1-800-827-6364) References: <9511130741.AA11137@rz1.rz.hs-bremen.de> In-Reply-To: <9511130741.AA11137@rz1.rz.hs-bremen.de> X-Mozilla-Status: 0011 Content-Length: 691 Good Morning; Steve Ediken and I talked about a much simpler version of your self flying kite idea about 15 years ago. Our idea (pre PC's) was to put a heavy battery in a metal box, an RC receiver, strong servos and a disguised antenna sitting on a beach. Two flying lines would come out of the box - go to a stunt kite - and "fly" the kite. Two small manekin forearms and hands (around handles) would come out of the box too. The pilot would sit (hidden) in his car, flying the kite... the idea was cut short by his very untimely passing from a kite accident. The whole idea was to be a giggle. I love what you're doing. Great stuff! All the best. Bill Tyrrell From ???@??? Sun Nov 19 12:52:08 1995 Received: from epsilon.qmw.ac.uk by rz1.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA08693; Thu, 16 Nov 1995 15:30:44 +0100 Received: from holly.elec.qmw.ac.uk by epsilon.qmw.ac.uk with SMTP-DNS (PP) id <02634-0@epsilon.qmw.ac.uk>; Wed, 15 Nov 1995 15:15:33 +0000 Received: by holly.elec.qmw.ac.uk (smail) id ; Wed, 15 Nov 95 15:09 GMT Date: Wed, 15 Nov 1995 15:09:26 +0000 (GMT) From: Andrew Hawken Subject: Re: Autopilot for Revolution (AutoRev) To: buchholz@fbm.hs-bremen.de Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Mozilla-Status: 0011 Content-Length: 2730 In article <9511130741.AA11137@rz1.rz.hs-bremen.de> you wrote: : Dear friends : Five weeks ago they gave me this job as a professor for flight control : at the Hochschule Bremen, Germany. One of the advantages of such a : business is the freedom to choose your own research topic: I will : invest my scientific energy into the modelling, simulation and : control of kites. Firstly, congratulations on your Professorship. I am genuinly pleased at the creation of your post and I am sure that rec.kites as a whole will try to give you its best. It is an important stepping stone in the evoltion of kites from kids toys to, ummm, adult toys :-) As for applications of the autorevrev. Here are some disorganised thoughts. They may just be usefull. 1. What is the advantage over a stable single line kite. In order for the autorev to be usefull, it must be treated as a mobile platform which an adjustable attitude compared to the relative wind. With a light payload, it can be manouvered away from the tow point, but hovered just above ground. 2. With computer winch control, rather than quad handles, you have a machine that has a true 3d envelope rather than the rev's 2D surface. 3. It can lift small payloads or be used to get direct feedback of wind conditions. 4. Would it be possible to use a rev as a power source, with the simulator to keep it in the sky. A big kite on long lines could be used to drive some sort of generator. 5. The kite part of the device would contain no onboard electronics (enless video feedback proves impossible). There may therefore be environments where a passive device such as a kite is acceptable but an active one (small hellicopter for instance) is not. I am not sure what its roll would be in identifying polutants, but I do feel that there is a niche where accurate scanning would be usefull. Given a system capable of 3d movement, mounted on a platform with a horisontal degree of freedom, it would be possible to accuratly traverse a large area of ground, without any disturbance. Could be usefull for archealogical or forensic work ? Tell me, are you going to be able to do any work on aerodynamics of kites, and find out how and why these things fly. Settle some of the questions about lift etc ? ____________________________Andrew Hawken______________________________ "However many ways there may be of being alive, it is certain that there are vastly more ways of being dead" R Dawkins. Home : 01895 420110 QMW : 0171 975 5542 AIIT : 01494 677045 Email : A.Hawken@QMW.AC.UK From ???@??? Sun Nov 19 12:52:10 1995 Received: from haymarket.ed.ac.uk by rz1.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA09611; Fri, 17 Nov 1995 10:44:05 +0100 Received: from castle.ed.ac.uk (castle.ed.ac.uk [129.215.128.23]) by haymarket.ed.ac.uk (8.6.10/8.6.12) with SMTP id JAA01019 for ; Fri, 17 Nov 1995 09:04:32 GMT From: biegler@castle.ed.ac.uk To: buchholz@fbm.hs-bremen.de Subject: Re: Autopilot for Revolution (AutoRev) Newsgroups: rec.kites References: <9511130741.AA11137@rz1.rz.hs-bremen.de> X-Newsreader: NN version 6.5.0 #3 (NOV) Date: Fri, 17 Nov 95 7:56:15 GMT Message-Id: <9511170756.aa23514@uk.ac.ed.castle> X-Mozilla-Status: 0011 Content-Length: 1059 In rec.kites you write: >To make the long story short: Right now I am looking for __serious__ >applications of such an AutoRev. I still have to find partners in >industry and ministry to support these research activities. If the thing can be scaled up (I assume you just need big enough servos) and used on soft kites, then large traction kites could provide sail assistance to cargo ships. Attached on the lee side they wouldn't give the heeling moment of a more conventional rig. Do a feasibility study on required kite size, cost, fuel savings etc, if that looks good point out that the only thing missing is something to fly the kite. Unless the algorithm is applicable to other problems, but best derived from kite control, I can't see anything other than maritime applications, simply because you need lots of empty space if you want to leave a kite flying unsupervised. However, the application you mentioned doesn't seem to require a steerable kite, only one that is as stable as possible. Robert Biegler r.biegler@ed.ac.uk From ???@??? Sun Nov 19 15:37:34 1995 To: daveculp@bdt.com (Dave Culp) From: buchholz@fbm.hs-bremen.de (Joerg J. Buchholz) Subject: Re: Autopilot for Revolution (AutoRev) Cc: Bcc: X-Attachments: X-Mozilla-Status: 0011 Content-Length: 313 Dear Dave That's incredible! How could you come up with so many terrific ideas just out of the blue?! Have you ever been working on that field or is it just your phenomenial(?) imagination? Thank you, thank you, thank you .... Joerg (I am sure they _will_ pay and you will get your AutoFlex!) From ???@??? Sun Nov 19 15:39:45 1995 To: Rodger Duffett From: buchholz@fbm.hs-bremen.de (Joerg J. Buchholz) Subject: Re: Autopilot for Revolution (AutoRev) Cc: Bcc: X-Attachments: X-Mozilla-Status: 0011 Content-Length: 367 Dear Rodger >I have long dreamed of a computerised autopilot to control a power-kite >on a boat. Only this would most likely not be a Rev type kite. I think >that it would be something like a super efficient, lightweight but very >strong glider tethered to the boat by a single 'string'! > How would you _control_ a single liner? Like a fighter? Joerg From ???@??? Sun Nov 19 15:50:56 1995 To: salanne@convex.csc.FI (Simo Salanne) From: buchholz@fbm.hs-bremen.de (Joerg J. Buchholz) Subject: Re: Autopilot for Revolution (AutoRev) Cc: Bcc: X-Attachments: X-Mozilla-Status: 0011 Content-Length: 159 Dear Simo >... Apparently we need to add a statement, that >robots (I don't mean Robos :-) don't qualify as licensed >competitors;-/ W H Y ? Joerg From ???@??? Sun Nov 19 16:35:28 1995 To: A.Hawken@qmw.ac.uk From: buchholz@fbm.hs-bremen.de (Joerg J. Buchholz) Subject: Re: Autopilot for Revolution (AutoRev) Cc: Bcc: X-Attachments: X-Mozilla-Status: 0011 Content-Length: 999 Dear Andrew Thank you very much for your nice words and your great thoughts and ideas on the AutoRev project. Especially the idea to use winch control for a _3D_ flight envelope will definitely be a good argument in the discussions to come. Aerodynamics: May be you read my post to rec.kites when there was a discussion on different aerodynamic kite models. On the other hand AutoRev will give me a perfect opportunity to really "identify" the parameters of a real-world kite. I will be able to input predefined control inputs (like steps, sine-sweeps, ...) into the system to be identified (the kite) and measure all the information about the kinematical and aerodynamical state of the kite from the ground via a video camera and a proper pattern recognition algorithm. These methods have been very successfully used in parameter estimation of aircraft and other vehicles over years and will definitely give us more insight into the aerodynamics of our kites too. Joerg From ???@??? Mon Nov 20 08:00:37 1995 Received: from convex.csc.fi by rz1.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA09536; Sun, 19 Nov 1995 16:59:56 +0100 Received: (from salanne@localhost) by convex.csc.fi (8.6.9/8.6.9+CSC-2.0) id SAA17844 for buchholz@fbm.hs-bremen.de; Sun, 19 Nov 1995 18:07:05 +0200 From: Simo Salanne Message-Id: <199511191607.SAA17844@convex.csc.fi> Subject: Re: Autopilot for Revolution (AutoRev) To: buchholz@fbm.hs-bremen.de (Joerg J. Buchholz) Date: Sun, 19 Nov 1995 18:07:04 +0200 (EET) In-Reply-To: <9511191444.AA09259@rz1.rz.hs-bremen.de> from "Joerg J. Buchholz" at Nov 19, 95 03:44:34 pm X-Mailer: ELM [version 2.4 PL24] Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mozilla-Status: 0011 Content-Length: 575 > > Dear Simo > > >... Apparently we need to add a statement, that > >robots (I don't mean Robos :-) don't qualify as licensed > >competitors;-/ > > W H Y ? > Joerg, do you know any sport, where men, women, juniors and seniors compete together with robots? In kiting we don't classify people by sex or age, but my humble opinion is that I prefer competing with humans. Actually, the present rules prevent the use of robots, "any ground props" are not allowed, except in freestyle. In chess there's demonstrations between man and machine... Simo From ???@??? Mon Nov 20 08:56:43 1995 To: Simo Salanne From: buchholz@fbm.hs-bremen.de (Joerg J. Buchholz) Subject: Re: Autopilot for Revolution (AutoRev) Cc: Bcc: X-Attachments: X-Mozilla-Status: 0011 Content-Length: 435 Dear Simo I hope you did not take my question too serious :-) I definitely do not want to let a robot actually _compete_ against humans! But as you said, there might be demonstrations ... And if you promise me not to tell anybody else, I might tell you that I am sure it will take years and years until my AutoRev will at least have a chance to perform aesthetic ballets. But it will be lotsa fun trying ...! Joerg From ???@??? Tue Nov 21 14:47:26 1995 Received: from spinne.ip.lu by rz1.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA09566; Mon, 20 Nov 1995 10:39:42 +0100 Message-Id: From: grg@ip.lu Subject: Re: Autopilot for Revolution (AutoRev) To: buchholz@fbm.hs-bremen.de (Joerg J. Buchholz) Date: Mon, 20 Nov 1995 11:03:19 +0100 (MET) In-Reply-To: <9511191442.AA07464@rz1.rz.hs-bremen.de> from "Joerg J. Buchholz" at Nov 19, 95 03:42:21 pm X-Mailer: ELM [version 2.4 PL23] Content-Type: text X-Mozilla-Status: 0011 Content-Length: 229 i'm quite proud of one of the shortest internet mail addresses; lu is for luxembourg; there is somewhere a table with all the countries fr it ch .. what about adding wheels to the robot and doing some 3d, indoor ! -grg- From ???@??? Tue Nov 21 14:47:25 1995 Received: from uctmail2.uct.ac.za by rz1.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA07917; Mon, 20 Nov 1995 09:53:19 +0100 Received: by uctmail2.uct.ac.za (Smail3.1.29.1 #14) id m0tHReY-0005pBC; Mon, 20 Nov 95 10:32 SAT Message-Id: Received: by uctmail2.uct.ac.za (Mort 2.0a) id 12231 from RAY; Mon Nov 20 10:32:37 1995 From: "Rodger Duffett" Organization: UCT To: buchholz@fbm.hs-bremen.de (Joerg J. Buchholz) Date: Mon, 20 Nov 1995 10:40:20 SAST-2 Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT Subject: Re: Autopilot for Revolution (AutoRev) Priority: normal X-Mailer: Pegasus Mail for Windows (v2.01) X-Mozilla-Status: 0011 Content-Length: 555 Hello Joerg, > How would you _control_ a single liner? Like a fighter? It would have flight control surfaces like a conventional glider. These would be controlled by radio - perhaps not as interesting a problem as that modelled by your revsim! Cheerio __________________________________________________________________ Rodger Duffett Internet Address: rodger@ray.uct.ac.za Dept. Radiation Oncology Telephone : +2721 404 3135 Radiobiology Section University of Cape Town Groote Schuur Hospital South Africa From ???@??? Tue Nov 21 14:50:01 1995 To: grg@ip.lu From: buchholz@fbm.hs-bremen.de (Joerg J. Buchholz) Subject: Re: Autopilot for Revolution (AutoRev) Cc: Bcc: X-Attachments: X-Mozilla-Status: 0011 Content-Length: 112 >what about adding wheels to the robot and doing some >3d, indoor ! >-grg- > > Absolutely fantastic idea! From ???@??? Sat Dec 02 14:53:28 1995 Received: from bdt.com by rz1.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA09963; Thu, 30 Nov 1995 02:16:28 +0100 Received: by bdt.bdt.com (/\oo/\ Smail3.1.29.1 #29.6) id ; Wed, 29 Nov 95 16:59 PST Date: Wed, 29 Nov 95 17:23 PST Received: from noident@daveculp.dial-up.bdt.com(204.182.11.161) by bdt via smap (V1.3bdt) id smaa23859; Wed Nov 29 16:58:43 1995 Message-Id: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: buchholz@rz1.rz.hs-bremen.de From: daveculp@bdt.com (Dave Culp) Subject: Re: Autopilot for Revolution (AutoRev) X-Mozilla-Status: 0011 Content-Length: 1506 >That's incredible! How could you come up with so many terrific ideas just >out of the blue?! Have you ever been working on that field or is it >just your phenomenial(?) imagination? > >Thank you, thank you, thank you .... I have given it a bit of thought... You see, I've lived and breathed kites and kite traction for many years (my wife will be the first to assure you that most everything else; job, family, etc. have been secondary.) Once you get into it further, I'd be happy to trouble-shoot, and/or bring you additional ideas. I'm very excited about someone finally doing this. I first discussed it with friends at a speedsailing meet in 1986... You should contact a couple of people in Europe. Theo Schmidt, in Switzerland gave me the scanning laser idea. Keith Stewart, of England worked for years trying to exploit commercial and military uses for (non-autopiloted) kites, back in the 80's. I'll look up their addresses for you. >(I am sure they _will_ pay and you will get your AutoFlex!) Could this actually happen? Of course, you'll need a beta testing site... Keep me in the loop! ############################################################################# Dave Culp Speedsailing | daveculp@bdt.com 312 Flaming Oak Drive | http://www.bdt.com:80/home/daveculp/speedsail.html Pleasant Hill, CA 94523 | (NEW Website) USA | (510) 933.9300 x 308 ############################################################################ # From - Wed Apr 10 10:21:31 1996 Newsgroups: rec.kites Path: gina.zfn.uni-bremen.de!news.dfn.de!news.rwth-aachen.de!news.ruhr-uni-bochum.de!news.uni-stuttgart.de!uni-regensburg.de!lrz-muenchen.de!informatik.tu-muenchen.de!Germany.EU.net!howland.reston.ans.net!psinntp!psinntp!psinntp!psinntp!se01!usenet From: Bob Neitzke Subject: Re: Laser-guided kite. Was Re: Bizarre kite laws X-Nntp-Posting-Host: neitzke Content-Type: text/plain; charset=us-ascii Message-ID: <31693368.7D57@elk.miles.com> Sender: usenet@se01.elk.miles.com (Usenet News) Content-Transfer-Encoding: 7bit Organization: Diagnostics QA/RA References: <694.6664T836T238@dialnet.net> <723.6666T117T351@dialnet.net> <4k624r$ap6@zinc.compulink.co.uk> Mime-Version: 1.0 Date: Mon, 8 Apr 1996 15:40:24 GMT X-Mailer: Mozilla 2.0 (Win16; I) Lines: 58 Mark de Roussier wrote: > handling capability ! The trick, IMHO, will certainly be in getting a > good model of the kites behavior i.e. As long as we are considering a computer to do the flying , why not let it ( the computer ) learn how to fly the kite. Basic moves for starters and observations of reactions followed by learned (predictive) responses. This is how we all learned to fly in the first place. Then afterwards we do whatever is required outside of the learned characteristics of each kite to accomodate the situation/conditions. I would lean more heavily on the feedback approach rather than attempting to know all variables and issue a fixed commands to achieve response. The servo control idea was acceptable at first but on further examination it appears that what is will required is a fast wind and unwind mechanism to imitate the forward and backward movement of a human kite flier. With this type of device the computer should be able to do serious stalls via fast unwind or even 360s while winding in and swiveling around the compass. This mechanism, too, would be a servo, however, I was thinking of simple radio control style units. Further, lots of power will be needed because simple gearing will slow down response. I suppose lots of batteries will assist in anchoring this whole "wondement". Alternative 1: Further, I must ask. Why use the laser at all? Vision based systems are fairly common these days. Just point a video camera at the kite flying window, identify / isolate the kite image and include the positional data in the feedaback loop. Video techniques are used for lots of production line quality control systems. Alternatice 2: Another radical approach would be to simply monitor line angle and tension. Line angle would have to be in both axes ( up down and left right). The idea being that knowing the angle of the lines in relationship to each other and other fixed references would yeild kite position, direction and speed, through angular velocity. No video or laser problems with this approach. HiRez angular resolvers would be a necessity however. It's always best to exhaust/analyze all alternatives before acutally beginning a project. It's hoped that the above notions will prompt further creative approaches. This project, though exciting, makes me wonder if what we really need it not a robot to assist us when we "blow it" and dump the kite with line wraps et al. Just send "Robby" the kitestraightening robot out to pick up the kite, unwind the wraps and set it back up for us. Of course I would never need this kind of assistance because I never dump my kites ;-} And I always ... Stay outa da treez Bob Neitzke birdofplay@aol.com http://memnbers.aol.com/birdofplay From - Wed Apr 10 10:47:52 1996 Path: gina.zfn.uni-bremen.de!informatik.uni-bremen.de!cs.tu-berlin.de!zib-berlin.de!uni-paderborn.de!golden-gate.owl.de!fu-berlin.de!news.belwue.de!news.uni-stuttgart.de!uni-regensburg.de!lrz-muenchen.de!informatik.tu-muenchen.de!Germany.EU.net!howland.reston.ans.net!vixen.cso.uiuc.edu!newsfeed.internetmci.com!in2.uu.net!zdc!szdc!news From: rasmith@magicnet.net (Roger Smith) Newsgroups: rec.kites Subject: Re: Laser-guided kite. Was Re: Bizarre kite laws Date: Mon, 08 Apr 1996 22:53:54 GMT Organization: Zippo Lines: 36 Message-ID: <316997f7.356595499@snews.zippo.com> References: <694.6664T836T238@dialnet.net> <723.6666T117T351@dialnet.net> NNTP-Posting-Host: fwl.orlando.veda.com X-Newsreader: Forte Agent .99d/32.182 >> >It's not so far fetched... You put a reflector on each kite tip >> >(bicycle-type will do), and set your ex-supermarket scanning radar to scan >> Impossible to do with that equipment. >I *know* that a supermarket scanner *off the shelf* won't work! It's the >concept I was talking about ... I suspect that it might be easier and less expensive to use a TV camera as the sensor. Using edge-detection and other pretty common algorithms on the image to determine the kite's orientation and position shouldn't be too difficult. The company I work for (Veda, Inc.) is using a system developed by another company called Biomechanics. The system converts the motion of people into 3D computer data using small reflectors positioned at key points on the person - it's a lot less restrictive than the older tethered systems. Video cameras are positioned around the "studio" and Biomechanic's software tracks each of the dots to drive the animation or analysis. It's pretty cool. If you've seen Batman Forever, you've seen what it can do. The scene where Batman falls/flys from the building, lands and walks off was all computer graphics which were done using this system. So, anyway, that's what made me think of using a TV camera (or two?) for your application. It wouldn't be as complicated as the system we use. I think with a single camera setup, some distinct markings on the kite, a frame grabber in a laptop PC, some hardware, and some fancy software you could, uhh, make this idea fly. :-) -- Roger * Give Kids the World: http://www.america.com/~dcop/tudlp/gktw.html From - Wed Apr 10 15:09:31 1996 Received: from fwl.orlando.veda.com by hermes.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA11562; Wed, 10 Apr 1996 14:45:23 +0200 Received: by fw1.orlando.veda.com; id IAA17356; Wed, 10 Apr 1996 08:58:46 -0400 Received: from dopey.orlando.veda.com(192.168.97.25) by fw1.orlando.veda.com via smap (V3.1) id xma017354; Wed, 10 Apr 96 08:58:45 -0400 From: rasmith@magicnet.net (Roger Smith) To: "Jvrg J. Buchholz" Subject: Re: Laser-guided kite. Was Re: Bizarre kite laws Date: Wed, 10 Apr 1996 12:56:31 GMT Message-Id: <316baee5.75797456@mailhost.magicnet.net> References: <694.6664T836T238@dialnet.net> <723.6666T117T351@dialnet.net> <316997f7.356595499@snews.zippo.com> <316B75A3.660B@fbm.hs-bremen.de> In-Reply-To: <316B75A3.660B@fbm.hs-bremen.de> X-Mailer: Forte Agent .99d/32.182 Status: U X-Mozilla-Status: 0011 Content-Length: 1275 >It looks like there has been a thread about a computer controlled kite >in rec.kites. I think that thread is just starting. One gentleman was talking about using a laser scanner to track the position of a kite. Someone else questioned that idea saying that it would be difficult. Then I replied that I think using a TV camera as a sensor would be a lot easier. I guess you saw my message. That's about as far as the discussion has progressed so far. >Some of you might know that about two years ago I gave birth to a real-time >pilot-in-the-loop simulator of a Revolution kite by the name of RevSim. >My next two years project will now be a Revolution autopilot (AutoRev), >i.e. a box in the grass controlling all four lines of a rev. >It will use a video camera and pattern recognition algorithms to receive >information about the state (attitude and position) of the kite and >utilize the mathematical model of RevSim for a model following control >algorithm in the control computer. That sounds like a very interesting (and challenging) project! I hope you are able to find someone to support it. Please keep us in rec.kites informed of your progress. -- Roger * Give Kids the World: http://www.america.com/~dcop/tudlp/gktw.html From - Thu Apr 11 08:19:17 1996 Received: from uu7.psi.com by hermes.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA13531; Wed, 10 Apr 1996 17:09:00 +0200 Received: from se01.UUCP by uu7.psi.com (5.65b/4.0.940727-PSI/PSINet) via UUCP; id AA18687 for ; Wed, 10 Apr 96 10:59:19 -0400 Received: from neitzke.elk.bayer.com by se01.elk.miles.com (4.1/96010800) id AA26299; Wed, 10 Apr 96 09:55:49 EST Message-Id: <316BE8AE.59A9@elk.miles.com> Date: Wed, 10 Apr 1996 09:58:22 -0700 From: Bob Neitzke Organization: Diagnostics QA/RA X-Mailer: Mozilla 2.0 (Win16; I) Mime-Version: 1.0 To: "Jvrg J. Buchholz" Subject: Re: Laser-guided kite. Was Re: Bizarre kite laws References: <694.6664T836T238@dialnet.net> <723.6666T117T351@dialnet.net> <4k624r$ap6@zinc.compulink.co.uk> <31693368.7D57@elk.miles.com> <316B728F.7EA5@fbm.hs-bremen.de> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mozilla-Status: 0011 Content-Length: 4343 Jeorg J. Buchholz wrote: > > Dear Bob > > You were talking about a computer controlled kite in the > above mentioned post to rec.kites. I was merely postulating some ideas for discussion purposes. It's fun to speculate about the possibilities. I did not save the rec.kites discussions, unfortuantely. Are you the person who wrote the REV simulation program ? > Could you please be so kind and mail me some of the discussion > on that topic? Who was involved in that discussion? The discussion has been limited to scanning laser capabilities untill I suggested some alternatives. Questions were raised regarding the capabilities, range, power etc. of certain "off the shelf" lasers. Again I believe everybody was "shooting from the hip" I'm not convinced as to the accuracy of all the comments. > > Five weeks ago they gave me this job as a professor for flight control > at the Hochschule Bremen, Germany. One of the advantages of such a > business is the freedom to choose your own research topic: I will > invest my scientific energy into the modelling, simulation and > control of kites. Talk about a cool job. You must have done wellin you academics or know the dean very well ;-) > > Some of you might know that about two years ago I gave birth to a real-time > pilot-in-the-loop simulator of a Revolution kite by the name of RevSim. OK that answers my question above ... I thought so ! > My next two years project will now be a Revolution autopilot (AutoRev), > i.e. a box in the grass controlling all four lines of a rev. > It will use a video camera and pattern recognition algorithms to receive > information about the state (attitude and position) of the kite and > utilize the mathematical model of RevSim for a model following control > algorithm in the control computer. OK ... this might be where I got recent, posted the idea from. This seed more logical to me, as opposed to the laser targeting approach. Another group in my company used photo imaging to do QA and final release of LCD display products. The system ran the machine and "watched" the display for correct results. Other depatements have used video techniques for "high Speed" mechanism studies and certain fabric/fluid wicking studies. > But - I believe that such a project could be a perfect opportunity > to promote the colorful idea of kiting in those gray academic areas > of science and research. Last week I gave a lecture on RevSim and > the mathematical model behind it at an international symposium on > real-time simulation in Delft. You should have seen those scientists > and managers, who usually talk about A320 and space simulators and > investments of millions of dollars, suddenly becoming interested in > the fascinating world of rev flying (I showed part of the Revolution > teaching video). A good kick in the seat of the pants usually gets results !!! Good for you !!! I'm currently stuck with how you will mechanize the lines. I believe that you will need high speed wind/unwind mechanics in order to simulate the kite pilot stepping forward or backward. This will also allow for unlimited line length, with in certain constraints, i.e. line elasticity, wind angle on lines kite size ( stuff like that ). > > I was thinking for example of some environmental applications to use > an autonomously flying kite based on a ship for the detection of > maritime waste (oil, ...). I'm not sure regarding this application. It would be cheaper than using aircraft but it's a whole lot slower. > > Could you help me out with other AutoRev applications that might > persuade the right people ... 1. Beach guard / Life Guard applications a. Watching for marine life ( sharks too. problem is most winds, at shorelines, blow in not out. 2. Forrest observation from existing towers but with new, higher, vantage point. 3. Moving target for milirtary exercises. 4. Airborne listening post for environmental studies involving animal habitat and behavior. Good luck and keep me / us posted on how things are going. Best regards, Bob -- Bob Neitzke CAD / MGR, Bayer Corporation Elkhart Site work 219-262-7023 neitzke@se01.elk.bayer.com home 616-663-2833 birdofplay@aol.com http://members.aol.com/birdofplay From - Thu Apr 11 08:19:17 1996 Received: from emout10.mx.aol.com by hermes.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA11352; Thu, 11 Apr 1996 05:13:44 +0200 Received: by emout10.mail.aol.com (8.6.12/8.6.12) id XAA13355 for buchholz@fbm.hs-bremen.de; Wed, 10 Apr 1996 23:25:50 -0400 Date: Wed, 10 Apr 1996 23:25:50 -0400 From: BirdofPlay@aol.com Message-Id: <960410232548_510815470@emout10.mail.aol.com> To: buchholz@fbm.hs-bremen.de Subject: Re: Laser-guided kite. X-Mozilla-Status: 0011 Content-Length: 996 Joerg J. Buchholz >Dear Bob > >You were talking about a computer controlled kite in the >above mentioned post to rec.kites. Just a reply from the other provider to follow up on the reply I authored from my "at work" news and mail readers. After an entire day of allowing your ideas, and mine, to steep in the darker regions of the psyche I still comming up short on alternative ideas for the automated kite. Short of the more crass enterprises i.e. as an advertising platform / gimmick I feel stumped as to creative uses of the envisioned device. Kites themselves are not measureably usefull even though they provide great uplift to the soul and greater relief to the eye. You might consider funding from the Arts rather than the Sciences. They understand fun and expression without being pressed for directly measurable utility. Again ... Good luck on your quest, Best Regards Bob Neitzke birdofplay@aol.com http://members.aol.com/birdofplay From - Thu Apr 11 08:33:19 1996 Message-ID: <316CA7AF.34A0@fbm.hs-bremen.de> Date: Thu, 11 Apr 1996 08:33:19 +0200 From: "Jörg J. Buchholz" Organization: Hochschule Bremen X-Mailer: Mozilla 2.01 (Win95; I) MIME-Version: 1.0 To: Bob Neitzke Subject: Re: Laser-guided kite. Was Re: Bizarre kite laws References: <694.6664T836T238@dialnet.net> <723.6666T117T351@dialnet.net> <4k624r$ap6@zinc.compulink.co.uk> <31693368.7D57@elk.miles.com> <316B728F.7EA5@fbm.hs-bremen.de> <316BE8AE.59A9@elk.miles.com> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mozilla-Status: 0011 Content-Length: 827 Bob Neitzke wrote: > > I'm currently stuck with how you will mechanize the lines. I believe > that you will need high speed wind/unwind mechanics in order to > simulate the kite pilot stepping forward or backward. This will also > allow for unlimited line length, with in certain constraints, > i.e. line elasticity, wind angle on lines kite size ( stuff like that ). Yes this is one of the more interesting challenges (not to say, my biggest problem). I really need some high speed high torque motors. Right now I'm thinking about a combination of one slower strong motors, winding up all four lines and four fast hydraulic cylinders to emulate the handle motion. Thanks alot four your help and your ideas. I like the environmental studies ... Joerg -- buchholz@fbm.hs-bremen.de (Jörg J. Buchholz) From - Thu Apr 11 08:56:08 1996 Message-ID: <316CAD08.4269@fbm.hs-bremen.de> Date: Thu, 11 Apr 1996 08:56:08 +0200 From: "Jörg J. Buchholz" Organization: Hochschule Bremen X-Mailer: Mozilla 2.01 (Win95; I) MIME-Version: 1.0 To: BirdofPlay@aol.com Subject: Re: Laser-guided kite. References: <960410232548_510815470@emout10.mail.aol.com> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mozilla-Status: 0011 Content-Length: 281 BirdofPlay@aol.com wrote: > You might consider funding from the Arts rather than the Sciences. > They understand fun and expression without being pressed for directly > measurable utility. Like that idea ... Joerg -- buchholz@fbm.hs-bremen.de (Jörg J. Buchholz) From - Thu Apr 11 16:51:37 1996 Received: from uu7.psi.com by hermes.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA10000; Thu, 11 Apr 1996 16:21:51 +0200 Received: from se01.UUCP by uu7.psi.com (5.65b/4.0.940727-PSI/PSINet) via UUCP; id AA01671 for ; Thu, 11 Apr 96 10:14:26 -0400 Received: from neitzke.elk.bayer.com by se01.elk.miles.com (4.1/96010800) id AA11039; Thu, 11 Apr 96 09:13:39 EST Message-Id: <316D304F.1B57@elk.miles.com> Date: Thu, 11 Apr 1996 09:16:15 -0700 From: Bob Neitzke Organization: Diagnostics QA/RA X-Mailer: Mozilla 2.0 (Win16; I) Mime-Version: 1.0 To: buchholz@fbm.hs-bremen.de Subject: robo-kite References: <9604111316.AF26135@uu7.psi.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Status: U X-Mozilla-Status: 0001 Content-Length: 1383 > Jvrg J. Buchholz wrote: > > > > > Yes this is one of the more interesting challenges (not to say, my biggest > > problem). I really need some high speed high torque motors. > > Right now I'm thinking about a combination of one slower strong motors, > > winding up all four lines and four fast hydraulic cylinders to emulate the > > handle motion. Two , blended , mechanisms sounds good to me - but more work for the programmer and more difficulty supplying power to each set of "muscles". This would limit ultimate transportablility and therefore utility. It's tough to imitate human capabilities sometimes. Another idea... Use for Site Survey. When it's too windy to aeriel survey with small aircraft or helicopter ( even models ) then drag out the robokite and hook up ( downlink ) the remote video cameras. BTW I observed that another poster had recommended to use of video positioning lately. I concure. Back to work for me ... Best regards, -- Bob Neitzke CAD / MGR, Bayer Corporation Elkhart Site work 219-262-7023 neitzke@se01.elk.bayer.com home 616-663-2833 birdofplay@aol.com http://members.aol.com/birdofplay -- Bob Neitzke CAD / MGR, Bayer Corporation Elkhart Site work 219-262-7023 neitzke@se01.elk.bayer.com home 616-663-2833 birdofplay@aol.com http://members.aol.com/birdofplay From - Tue Apr 16 15:06:23 1996 Path: gina.zfn.uni-bremen.de!informatik.uni-bremen.de!cs.tu-berlin.de!uni-erlangen.de!kue!lrz-muenchen.de!informatik.tu-muenchen.de!Germany.EU.net!Frankfurt.Germany.EU.net!news.maz.net!ins.net!blackbush.xlink.net!news.ecrc.de!news.sprintlink.net!new-news.sprintlink.net!newsfeed.internetmci.com!swrinde!cs.utexas.edu!howland.reston.ans.net!news-e2a.gnn.com!newstf01.news.aol.com!newsbf02.news.aol.com!not-for-mail From: birdofplay@aol.com (BirdofPlay) Newsgroups: rec.kites Subject: Re: Laser-guided kite. Was Re: Bizarre kite laws Date: 13 Apr 1996 11:59:04 -0400 Organization: America Online, Inc. (1-800-827-6364) Lines: 17 Sender: root@newsbf02.news.aol.com Message-ID: <4koj08$e8f@newsbf02.news.aol.com> References: <4kkjha$76p@newsbf02.news.aol.com> NNTP-Posting-Host: newsbf02.mail.aol.com X-Newsreader: AOL Offline Reader Just another thought ... Maybe we should change this subject line to ... Robo Kite Start out fresh, gather some more support, ellicit more ideas, build up some steam. Till someone actually gets some funding all we can do is talk about it. A lot of refinement can be accomplished without spending a penny. Fun doesn't have to be expensive. What say you ??? Bob From - Tue Apr 16 15:26:04 1996 Message-ID: <31739FEC.2D2C@fbm.hs-bremen.de> Date: Tue, 16 Apr 1996 15:26:04 +0200 From: "Jörg J. Buchholz" Organization: Hochschule Bremen X-Mailer: Mozilla 2.01 (Win95; I) MIME-Version: 1.0 Newsgroups: rec.kites Subject: AutoRev. Was Re: Laser-guided kite References: <4kkjha$76p@newsbf02.news.aol.com> <4koj08$e8f@newsbf02.news.aol.com> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mozilla-Status: 0001 Content-Length: 3878 BirdofPlay wrote: > > Just another thought ... > > Maybe we should change this subject line to ... > > Robo Kite > > Start out fresh, gather some more support, ellicit more > ideas, build up some steam. > > Till someone actually gets some funding all we can do > is talk about it. A lot of refinement can be accomplished > without spending a penny. > Fun doesn't have to be expensive. > > What say you ??? > > Bob It's time to _do_ it! They gave me a few thousand bucks and asked me to start working on the revolution auto pilot project (AutoRev). So I would like to share some of my current thoughts, ideas, and questions with you: During the next two years I will work on a Revolution auto pilot, which will be a box in the grass controlling all four lines of a rev. It will use a video camera and pattern recognition algorithms to receive information about the state (attitude and position) of the kite and utilize the mathematical model of RevSim (the Revolution Simulator) for a model following control algorithm in the control computer. The project divides into three main tasks: o Develop a computer controlled mechanical device to manipulate all four lines. The most simple way to do this would be four motors (electric or hydraulic), one for each line. These motors have to be fast _and_ strong. An alternative would be one strong slow motor, winding up all four lines (for 3D-flying) and four fast hydraulic cylinders to emulate the handle motion. What about a platform rotating about its vertical axis for indoor flying? o Use a video camera aimed at the rev to take images of the kite in every sample step. Do some pattern recognition with these images in the control computer (mainly edge detection), in order to come up with the three degrees of freedom of the kite, which are the azimuth angle, the elevation angle (representing the position of the kite on the flyable quarter dome), and the rolling angle about the line axis, representing the attitude of the kite. o These three angles are the values to be controlled by the control algorithm. A simple PID controller will definitely not do that job. Think about all the complicated handle coordinations you had to _learn_ before you could at least stabilize the rev. One way to use all that information you have about the characteristics of the plant you want to control is to implement a mathematical model of the plant in the feedforward controller of a model following control algorithm. Fortunately you have a sophisticated mathematical model of a revolution kite by the name of RevSim. So, by _simulating_ the kite in real time in the control loop AutoRev kind of "knows" (just like he human pilot) which will be the correct line motion for a desired maneuver. Having just started, there are a few questions I would like to discuss with you: o What is the mean force per line of a Rev II? What is the maximum force? o What is the maximum handle rotation speed (deg/sec) a human pilot has to use in order to control the kite? What about the maximum handle rotation acceleration (deg/sec/sec)? Is it important? o Hydraulic or electric motors? o Power supply (batteries?, ac/dc?)? I don't have a car. o Is a simple video camera (Hi8?) with the usual video-in-card for a PC sufficient for the resolution of let's say 10 degrees of rolling angle? At which line length? o What would be the best color of the kite for the pattern recognition? (blue sky?, white clouds?) o Is a standard PC (P90, programmed in C) fast enough to do pattern recognition and model following control in real time? Any idea? Joerg -- buchholz@fbm.hs-bremen.de (Jörg J. Buchholz) From - Wed Apr 17 06:43:19 1996 Received: from bfkvax.fm.bs.dlr.de by hermes.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA13525; Wed, 17 Apr 1996 04:43:46 +0200 Received: by bfkvax.fm.bs.dlr.de (UCX V3.3-7A); Wed, 17 Apr 1996 04:55:31 +0200 Received: from mail04.mail.aol.com (mail04.mail.aol.com [152.163.172.53]) by dv.ba.dlr.de (8.7.1/8.7.1) with SMTP id EAA32425 for ; Wed, 17 Apr 1996 04:55:58 +0200 From: BirdofPlay@aol.com Received: by mail04.mail.aol.com (8.6.12/8.6.12) id WAA04239 for joerg.buchholz@dlr.de; Tue, 16 Apr 1996 22:55:27 -0400 Date: Tue, 16 Apr 1996 22:55:27 -0400 Message-Id: <960416225516_192603113@mail04> To: fk2m@bfkvax.fm.bs.dlr.de Subject: Re: [Fwd: Re: Laser-guided kite. Was Re: Bizarre kite laws] X-Mozilla-Status: 0011 Content-Length: 493 Jeorg, We have been having problems getting e-mail thru to you ??? This, my latest post, came back - here it is again ... >My latest idea was to put some transducers in the lines >and take tension data as a kite was flown and sync it >with video taken simultaneously. >This would help "size" the mechanical requirements and >possibly could be used to "teach" the computer to fly. Hope this gets thru !!! Bob Neitzke birdofplay@aol.com http://members.aol.com/birdofplay From - Tue Apr 16 15:06:23 1996 Received: from bfkvax.fm.bs.dlr.de by hermes.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA10183; Tue, 16 Apr 1996 12:56:44 +0200 Received: from mail.euronet.nl (mail.euronet.nl [194.134.0.67]) by dv.ba.dlr.de (8.7.1/8.7.1) with ESMTP id NAA09000 for ; Tue, 16 Apr 1996 13:08:30 +0200 Received: from [194.134.3.15] (p027.mas.euronet.nl [194.134.3.33]) by mail.euronet.nl (8.7.5/8.7.3) with SMTP id NAA19685 for ; Tue, 16 Apr 1996 13:08:14 +0200 (MET DST) X-Sender: janv@euronet.nl (Unverified) Message-Id: Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Date: Tue, 16 Apr 1996 13:04:40 +0200 To: fk2m@bfkvax.fm.bs.dlr.de From: janv@euronet.nl (Jan Vegt) Subject: Re: AutoRev Content-Transfer-Encoding: quoted-printable Status: U X-Mozilla-Status: 0011 Content-Length: 6975 Dear J=F6rg, I tried to reply to your posting in rec.kites but my e-mail bounced. Than= ks to Peter Peters I got this e-mail address, I hope this works. Here's my original message : I came across your very interesting (forwarded) message about AutoRev in the rec.kites newsgroup. My background is in library and information scie= nce - so I'm not particularly technical - but I've always (that's since my teens, I'm 35 right now ;-) been interested in kites in relation to wind energy generation. Actually the reason for reading the rec.kites newsgrou= p is my believe that kites could be a cheap and effective way to harness wi= nd energy. Partly that's already proven technology (i.e. "power kiting" for ships, b= uggies etc.) My interest lies particularly with a type of wind generator known i= n Dutch as a "ringbaanturbine" ( a German equivalent might be very close ; RingbahnTurbine maybe? ;-) This type of windturbine goes back to the 1920= 's when it was devised and tried with Flettner rotors. I know there is some modern research done into this type of turbine (in the US and the SU), bu= t I'm not familiar with any literature apart from a Dutch thesis done somewhere in the 80s. The basic idea of this type of turbine is an endless track with 'buggies' riding on this track, the carts are of course propelled by windenergy. I think kites could be a very attractive (cheap and mechanically very sim= ple) device to power these 'buggies'. In fact as I see it, the main problem being the automatic tracking and control, since preferably human 'drivers= ' could be eliminated in this plan ;-) I'm biased of course but to me it seems a very good - real world - application for your type of research .... I've just started out my own company (software for the Apple Mac) so unfortunately I don't have the $30.000,- lying around. I may offer you some assistance though in doing some literature legwork (maybe compiling a literature report on this type of windturbine?) if you're interested. I'll be interested in your thoughts on this subject, Best regards, Jan Jan Vegt Karel Doormanlaan 52 3572 NE Utrecht The Netherlands ------------- From: birdofplay@aol.com (BirdofPlay) Newsgroups: rec.kites Subject: Re: Laser-guided kite. Was Re: Bizarre kite laws Date: 11 Apr 1996 23:43:38 -0400 Organization: America Online, Inc. (1-800-827-6364) Lines: 134 >>Mark de Roussier wrote: >>> handling capability ! The trick, IMHO, will certainly be in getting a >>> good model of the kites behavior i.e. >> Then Bob Neitzke wrote: >>As long as we are considering a computer to do the flying , why not >>let it ( the computer ) learn how to fly the kite. > Then Mark wrote: >I have considered this problem at some length. Now my head hurts. I >should learn not to think so hard and fly more kites... Bob concurs with similar head condition. But, this is the guy we should be talking directly to ... Joerg J. Buchholz buchholz@fbm.hs-bremen.de (J=96rg J. Buchholz) Forward your super neato techy ideas to him. Here is a message that he sent me recently. I believe that his news is screwed up so he e-mailed me re the thread. I have not saved any of the stuff re this thread so maybe someone else could help out here . Read his attachment , it's interesting. transcript follows .... Subj: Re: Laser-guided kite. Was Re: Bizarre kite laws Date: 96-04-10 06:32:46 EDT From: buchholz@fbm.hs-bremen.de (J=96rg J. Buchholz) To: neitzke@elk.miles.com (Bob Neitzke) CC: birdofplay@aol.com Some stuff deleted ... Since then I have not been able to follow the discussion on rec.kites. Could you please be so kind and mail me some of the discussion on that topic? Who was involved in that discussion? Thank you very much Joerg J. Buchholz -- buchholz@fbm.hs-bremen.de (J=96rg J. Buchholz) Attachment: Post to rec.kites Dear friends Five weeks ago they gave me this job as a professor for flight control at the Hochschule Bremen, Germany. One of the advantages of such a business is the freedom to choose your own research topic: I will invest my scientific energy into the modelling, simulation and control of kites. Some of you might know that about two years ago I gave birth to a real-time pilot-in-the-loop simulator of a Revolution kite by the name of RevSim. My next two years project will now be a Revolution autopilot (AutoRev), i.e. a box in the grass controlling all four lines of a rev. It will use a video camera and pattern recognition algorithms to receive information about the state (attitude and position) of the kite and utilize the mathematical model of RevSim for a model following control algorithm in the control computer. This will enable me to predefine a complete ballet in the computer (e.g. via Kiteflight) in advance and then let the autopilot win championships against human competitors. Isn't that a challenge? Will AutoRev ever do the Axel? I am totally aware of the fact that no kite flyer needs a kite flying machine. _We_ want to have fun with our kites, _we_ want to fly them ... So, don't take the paragraph above too seriously. But - I believe that such a project could be a perfect opportunity to promote the colorful idea of kiting in those gray academic areas of science and research. Last week I gave a lecture on RevSim and the mathematical model behind it at an international symposium on real-time simulation in Delft. You should have seen those scientists and managers, who usually talk about A320 and space simulators and investments of millions of dollars, suddenly becoming interested in the fascinating world of rev flying (I showed part of the Revolution teaching video). To make the long story short: Right now I am looking for __serious__ applications of such an AutoRev. I still have to find partners in industry and ministry to support these research activities. The Hochschule Bremen is not willing to fund some $30.000 just for the fun of it. I was thinking for example of some environmental applications to use an autonomously flying kite based on a ship for the detection of maritime waste (oil, ...). Could you help me out with other AutoRev applications that might persuade the right people ... Thank you very much -- buchholz@fbm.hs-bremen.de (J=96rg J. Buchholz) End Transcript ---------- So there you have it. I hope this thread gets even more interesting. As you just read J=96rg needs more ideas to justify the whole project. Lets get those skull caps off and let the breezes blow the cobwebs out. Nothing makes a project move faster than an "REAL" need. I read else where the the Puns are coming along nicely. My suggestions have already been e-mailed to J=96rg. OK folks lets give it a go ! Bob Neitzke birdofplay@aol.com http://members.aol.com/birdofplay From - Wed Apr 17 18:35:12 1996 Path: gina.zfn.uni-bremen.de!informatik.uni-bremen.de!cs.tu-berlin.de!uni-erlangen.de!kue!lrz-muenchen.de!informatik.tu-muenchen.de!Germany.EU.net!howland.reston.ans.net!plug.news.pipex.net!pipex!tank.news.pipex.net!pipex!lade.news.pipex.net!pipex!dsbc.icl.co.uk!jim From: jim@oasis.icl.co.uk (Jim Cheetham) Newsgroups: rec.kites Subject: Re: AutoRev. Was Re: Laser-guided kite Date: 17 Apr 1996 10:29:52 GMT Organization: ICL, Bracknell, UK Lines: 25 Message-ID: <4l2h70$2k7@eccles.dsbc.icl.co.uk> References: <31739FEC.2D2C@fbm.hs-bremen.de> Reply-To: jim@bra01.icl.co.uk NNTP-Posting-Host: vpm.bra01.icl.co.uk X-Newsreader: Tin 1.1 PL5 "Jörg J. Buchholz" (buchholz@fbm.hs-bremen.de) wrote: : It's time to _do_ it! : They gave me a few thousand bucks and asked me to start : working on the revolution auto pilot project (AutoRev). I tried to reply by email, but my mailer claims that it can't find fbm.hs-bremen.de ... Anyway, I was suggesting an email list for this project, with summaries in rec.kites every so often, once it gets going. That way, you can mix up kite flyers, robotics and electronics people without having a load of off-topic posts in any given newsgroup. If you need someone to host the list, I will be able to do that. I was also hoping to see your machine flying at festivals (which means, try to get the power requirements down one day, to something that can be battery powered, rather than requiring a generator ...) -- _____ ceci n'est ____ _ pas une _ _ email (__ __) o ______ ( __)( )_ ___ ___ _( )_( )_ ___ ______ (____) (_)(_)()(_) (____)(_)_)(__=)(__=) (_)_(_)_)(___)_(_)()(_) www.guernsey.net/~jim +44 1344 472537 From - Wed Apr 17 18:32:14 1996 Path: gina.zfn.uni-bremen.de!informatik.uni-bremen.de!cs.tu-berlin.de!zrz.TU-Berlin.DE!fu-berlin.de!news.belwue.de!news.uni-stuttgart.de!news.ruhr-uni-bochum.de!news.rhrz.uni-bonn.de!news.rwth-aachen.de!Aachen.Germany.EU.net!rmi.de!Germany.EU.net!howland.reston.ans.net!tank.news.pipex.net!pipex!lade.news.pipex.net!pipex!madge.co.uk!news From: mgibson@madge.com (Martin Gibson) Newsgroups: rec.kites Subject: Re: AutoRev. Was Re: Laser-guided kite Date: Tue, 16 Apr 1996 16:01:15 GMT Organization: Madge Networks Lines: 64 Message-ID: <4l0gfk$6ce@khyber.dev.madge.com> References: <4kkjha$76p@newsbf02.news.aol.com> <4koj08$e8f@newsbf02.news.aol.com> <31739FEC.2D2C@fbm.hs-bremen.de> Reply-To: mgibson@madge.com NNTP-Posting-Host: 194.32.220.165 X-Newsreader: Forte Free Agent 1.0.82 Jörg J. Buchholz" wrote: >During the next two years I will work on a Revolution auto pilot, >which will be a box in the grass controlling all four lines of a rev. >It will use a video camera and pattern recognition algorithms to receive >information about the state (attitude and position) of the kite and >utilize the mathematical model of RevSim (the Revolution Simulator) >for a model following control algorithm in the control computer. I have a couple of suggestions. I think your basic breakdown of the tasks is about right, but you may be able to simplify it further. >o Develop a computer controlled mechanical device to manipulate all > four lines. The most simple way to do this would be four motors > (electric or hydraulic), one for each line. These motors have to > be fast _and_ strong. An alternative would be one strong slow motor, > winding up all four lines (for 3D-flying) and four fast hydraulic > cylinders to emulate the handle motion. What about a platform > rotating about its vertical axis for indoor flying? I agree that you should start with the mechanical part. I suggest that your first step could be to duplicate the hand movements a human pilot makes to fly the kite. Effectively, this is fly by wire. I don't think you will have any problem matching the servo speeds to hand movements. If you've ever seen small industrial robots in action, you'll know what can be achieved if you do it right. I think you'll only need about 2-3Hz, but some safety margin would be wise. >o Use a video camera aimed at the rev to take images of the kite in > every sample step. Do some pattern recognition with these images > in the control computer (mainly edge detection), in order to come > up with the three degrees of freedom of the kite, which are the > azimuth angle, the elevation angle (representing the position > of the kite on the flyable quarter dome), and the rolling angle > about the line axis, representing the attitude of the kite. There is another, simpler way of determining the position & attitude of the kite. Apologies if this has been posted before; I'm fairly new to this thread, and an erratic reader of rec.kites. Measure the direction of the line "vector". This gives you the position of the kite. Do this for a couple of lines, and you know its orientation too. Admittedly this assumes the lines remain straight, but I think it's a reasonable assumption to start with, and you can always tune out line sag if you have to. I suggest an optical shaft encoder for each of the vertical & horizontal directions. No horrible pattern recognition required. >o These three angles are the values to be controlled by the control > algorithm. A simple PID controller will definitely not do that job. > Think about all the complicated handle coordinations you had to > _learn_ before you could at least stabilize the rev. I disagree. A fast PID controller can cope with things that humans can't attempt to do. There are some coordinations to be learned, but I don't think they're that complex. Humans can't touch controllers designed for specific purposes, but they're truly great at adaptive control. Get your parameters approximately right, and I believe it will work. Martin Gibson Madge Networks Ltd mgibson@madge.com From - Wed Apr 17 18:31:49 1996 Message-ID: <31751CF5.70C4@fbm.hs-bremen.de> Date: Wed, 17 Apr 1996 18:31:49 +0200 From: "Jörg J. Buchholz" Organization: Hochschule Bremen X-Mailer: Mozilla 2.01 (Win95; I) MIME-Version: 1.0 Newsgroups: rec.kites Subject: Re: AutoRev. Was Re: Laser-guided kite References: <4kkjha$76p@newsbf02.news.aol.com> <4koj08$e8f@newsbf02.news.aol.com> <31739FEC.2D2C@fbm.hs-bremen.de> <4l0gfk$6ce@khyber.dev.madge.com> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mozilla-Status: 0011 Content-Length: 2867 Martin Gibson wrote: > I agree that you should start with the mechanical part. I suggest > that your first step could be to duplicate the hand movements a human > pilot makes to fly the kite. Effectively, this is fly by wire. > I don't think you will have any problem matching the servo speeds to > hand movements. If you've ever seen small industrial robots in > action, you'll know what can be achieved if you do it right. > I think you'll only need about 2-3Hz, but some safety margin would be > wise. Controlling a rev does not only involve _hand_movements_ for the angles of attack of both sails but also _arm_movements_ (the two-liner style) for the generation of an angle of sideslip. The robot would have at least four motors and two joints. > There is another, simpler way of determining the position & attitude > of the kite. Apologies if this has been posted before; I'm fairly new > to this thread, and an erratic reader of rec.kites. > Measure the direction of the line "vector". This gives you the > position of the kite. Do this for a couple of lines, and you know its > orientation too. Admittedly this assumes the lines remain straight, > but I think it's a reasonable assumption to start with, and you can > always tune out line sag if you have to. I suggest an optical shaft > encoder for each of the vertical & horizontal directions. > No horrible pattern recognition required. I did look at the line angles while flying my rev in medium wind with light turbulences and tried to find a correlation between the rolling angle (attitude) of the kite and the line angles. The signal-to-noise ratio was not very encouraging. If the rolling angle is greater than 360 deg _no_ rolling angle information can be gained from the line angles at all, because the lines are crossed. > I disagree. A fast PID controller can cope with things that humans > can't attempt to do. > There are some coordinations to be learned, but I don't think they're > that complex. Humans can't touch controllers designed for specific > purposes, but they're truly great at adaptive control. > Get your parameters approximately right, and I believe it will work. A human pilot is a good pilot because he _knows_ his kite (car, a/c). The best way to implement knowledge about a plant into a control system is to use an "inverse model" of the plant in the feed forward path of the control system. A golden rule for control system design says:"Use all the information you have about the plant in the feed forward controller and let the feedback controller (PID controller) only compensate for disturbances and model uncertainties". I feel like this question should not be discussed in rec.kites ... :-) Thank you very much for your ideas. Joerg -- buchholz@fbm.hs-bremen.de (Jörg J. Buchholz) From - Wed Apr 17 15:15:18 1996 Message-ID: <3174EEE6.3C6B@fbm.hs-bremen.de> Date: Wed, 17 Apr 1996 15:15:18 +0200 From: "Jörg J. Buchholz" Organization: Hochschule Bremen X-Mailer: Mozilla 2.01 (Win95; I) MIME-Version: 1.0 To: Jan Vegt Subject: Re: AutoRev References: Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mozilla-Status: 0011 Content-Length: 630 Dear Jan Thank you very much for your ideas about using AutoRev for the control of a "ringbaanturbine". It really seems to be a very interesting plan to use wind energy in such a way. Assuming your implicit assent I will add your idea to my stack of possible AutoRev applications. On the other hand I already found some funding, so now I am eagerly starting the "real job", getting AutoRev to fly my revolutions, without worrying too much about actual applications. Nevertheless, thank you very much for your offer regarding the literature report. Joerg -- buchholz@fbm.hs-bremen.de (Jörg J. Buchholz) From - Wed Apr 17 07:01:05 1996 Message-ID: <31747B11.3FE9@fbm.hs-bremen.de> Date: Wed, 17 Apr 1996 07:01:05 +0200 From: "Jörg J. Buchholz" Organization: Hochschule Bremen X-Mailer: Mozilla 2.01 (Win95; I) MIME-Version: 1.0 To: BirdofPlay@aol.com Subject: Re: [Fwd: Re: Laser-guided kite. Was Re: Bizarre kite laws] References: <960416225516_192603113@mail04> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mozilla-Status: 0011 Content-Length: 1234 BirdofPlay@aol.com wrote: > > Jeorg, > > We have been having problems getting e-mail thru to you ??? > > This, my latest post, came back - here it is again ... > > >My latest idea was to put some transducers in the lines > >and take tension data as a kite was flown and sync it > >with video taken simultaneously. > >This would help "size" the mechanical requirements and > >possibly could be used to "teach" the computer to fly. > > Hope this gets thru !!! You made it. I really don't know what happened to your first mail. The guys from our computer center say they might have had some trouble with their mail server ... Your idea: Measuring the line forces during flight could indeed improve the control precision because if you feedback "inner" variables (forces, ...) as well, the controller can react faster. It does not have wait, until a disturbance shows up in the actual control variables (position, attitude). On the other hand, it might be worth trying without a cascade control, mainly to Keep It Small and Simple (KISS). Once again, thank you very much for your participation and your assistance with AutoRev. Joerg -- buchholz@fbm.hs-bremen.de (Jörg J. Buchholz) From - Wed Apr 17 20:34:00 1996 Message-ID: <31753998.4A75@fbm.hs-bremen.de> Date: Wed, 17 Apr 1996 20:34:00 +0200 From: "Jörg J. Buchholz" Organization: Hochschule Bremen X-Mailer: Mozilla 2.01 (Win95; I) MIME-Version: 1.0 Newsgroups: rec.kites Subject: Re: AutoRev References: <4kkjha$76p@newsbf02.news.aol.com> <4koj08$e8f@newsbf02.news.aol.com> <31739FEC.2D2C@fbm.hs-bremen.de> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mozilla-Status: 0011 Content-Length: 848 Andrew Beattie wrote: > I don't think that you want to look at the line movement in terms of distance. > Think more in terms of *tension* on the line. > > Build a system to monitor tension in a running line: > > /O\ > O / X \ O > -----------/ \---------- > > Where "O" are pulleys and X is a compression load-cell. > > And build a feedback loop, so that the motor pulls in line to increase > the pressure and lets it out to decrease it. You are definitely right. You can indeed utilize the line tension in a fast inner feedback loop of a cascade control system. On the other hand you still need the outer visual feedback loop because it is not the line tension you want to control but the position and the attitude of the kite. -- buchholz@fbm.hs-bremen.de (Jörg J. Buchholz) From - Thu Apr 18 07:08:02 1996 Message-ID: <3175CE32.17CD@fbm.hs-bremen.de> Date: Thu, 18 Apr 1996 07:08:02 +0200 From: "Jörg J. Buchholz" Organization: Hochschule Bremen X-Mailer: Mozilla 2.01 (Win95; I) MIME-Version: 1.0 Newsgroups: rec.kites Subject: Re: AutoRev. References: <4kkjha$76p@newsbf02.news.aol.com> <4koj08$e8f@newsbf02.news.aol.com> <31739FEC.2D2C@fbm.hs-bremen.de> <4l2jl3$5da@news.internetmci.com> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mozilla-Status: 0011 Content-Length: 481 4330291@mcimail.com wrote: > ... Personally, I'd approach it with a > single tethered line, with the servos operating control surfaces of some > sort. I don't believe monster servos operating the lines(ie: from the ground > as a person would) is the route to go ... I want to be able to let AutoRev fly *any* rev, not just a special version with servos and control surfaces. The machinery has to stay on the ground. -- buchholz@fbm.hs-bremen.de (Jörg J. Buchholz) From - Wed Apr 17 20:01:12 1996 Newsgroups: rec.kites Path: gina.zfn.uni-bremen.de!informatik.uni-bremen.de!cs.tu-berlin.de!uni-erlangen.de!kue!lrz-muenchen.de!informatik.tu-muenchen.de!Germany.EU.net!Frankfurt.Germany.EU.net!news.maz.net!ins.net!blackbush.xlink.net!news.ecrc.de!news.sprintlink.net!new-news.sprintlink.net!tank.news.pipex.net!pipex!lade.news.pipex.net!pipex!laura.pcug.co.uk!tug!andrew From: andrew@tug.com (Andrew Beattie) Subject: Re: AutoRev. Was Re: Laser-guided kite Organization: /usr/lib/news/organisation Date: Wed, 17 Apr 1996 05:57:41 GMT Message-ID: References: <4kkjha$76p@newsbf02.news.aol.com> <4koj08$e8f@newsbf02.news.aol.com> <31739FEC.2D2C@fbm.hs-bremen.de> Lines: 23 writes: >So I would like to share some of my current thoughts, ideas, >and questions with you: So here is a a thought of mine... I don't think that you want to look at the line movement in terms of distance. Think more in terms of *tension* on the line. Build a system to monitor tension in a running line: /O\ O / X \ O -----------/ \---------- Where "O" are pulleys and X is a compression load-cell. And build a feedback loop, so that the motor pulls in line to increase the pressure and lets it out to decrease it. Andrew -- The *real* bandwidth problem is the time it takes to ignore junk. From - Thu Apr 18 08:15:28 1996 Path: gina.zfn.uni-bremen.de!informatik.uni-bremen.de!cs.tu-berlin.de!fu-berlin.de!golden-gate.owl.de!uni-paderborn.de!news.rwth-aachen.de!Aachen.Germany.EU.net!rmi.de!Germany.EU.net!Frankfurt.Germany.EU.net!news.maz.net!ins.net!blackbush.xlink.net!howland.reston.ans.net!newsfeed.internetmci.com!news.sprintlink.net!new-news.sprintlink.net!rain.fr!oops.ip.lu!not-for-mail From: grg@oops.ip.lu (blabla) Newsgroups: rec.kites Subject: Re: AutoRev. Was Re: Laser-guided kite Date: 17 Apr 1996 12:54:46 +0200 Organization: Infopartners S.A. Luxembourg Lines: 79 Message-ID: <4l2ilm$ct0@oops.ip.lu> References: <4kkjha$76p@newsbf02.news.aol.com> <4koj08$e8f@newsbf02.news.aol.com> <31739FEC.2D2C@fbm.hs-bremen.de> NNTP-Posting-Host: oops.ip.lu X-Newsreader: TIN [version 1.2 PL2] Jörg J. Buchholz (buchholz@fbm.hs-bremen.de) wrote: : o Develop a computer controlled mechanical device to manipulate all : four lines. The most simple way to do this would be four motors : (electric or hydraulic), one for each line. These motors have to : be fast _and_ strong. An alternative would be one strong slow motor, : winding up all four lines (for 3D-flying) and four fast hydraulic : cylinders to emulate the handle motion. What about a platform : rotating about its vertical axis for indoor flying? will be the hardest part; i suggest that you should add to these mecanics a sensor capable of indicating the tension on each individual line, so that the machine gets some 'feeling'. : o These three angles are the values to be controlled by the control : algorithm. A simple PID controller will definitely not do that job. : Think about all the complicated handle coordinations you had to : _learn_ before you could at least stabilize the rev. One way to : use all that information you have about the characteristics of : the plant you want to control is to implement a mathematical : model of the plant in the feedforward controller of a model : following control algorithm. Fortunately you have a sophisticated : mathematical model of a revolution kite by the name of RevSim. : So, by _simulating_ the kite in real time in the control loop : AutoRev kind of "knows" (just like he human pilot) which will be : the correct line motion for a desired maneuver. another suggestion: as a human, you fly the rev by feeling (i do at least..), i.e. you do not try a manouevre which the kite is not prepared to follow, or by experience or by 'feeling'. To integrate these parameters into a model is IMHO only possible using fuzzy logic. i guess it will also be faster in calculus than any algorithmic programming. : Having just started, there are a few questions I would like to : discuss with you: : o What is the mean force per line of a Rev II? : What is the maximum force? depends on the wind no?, for 2 years i'm using the same 40kg lines up to 50km/h wind on a rev2 (spider line). : o What is the maximum handle rotation speed (deg/sec) a human : pilot has to use in order to control the kite? : What about the maximum handle rotation acceleration (deg/sec/sec)? : Is it important? maximum acceleration is required to bring the rev into trick positions. for normal flight they should not be no need. with for independantly controlled lines, you will have a complete new world available for tricks. real world revs have two lines on one handle, this you can simulate with only one motor: ---line----------------------/-\ |O| ---line----------------------\-/ the line going 3-4 times around the motor axis. then you need another motor to move above forward,backward : o Is a simple video camera (Hi8?) with the usual video-in-card : for a PC sufficient for the resolution of let's say 10 degrees : of rolling angle? At which line length? why not using the lines as position indicators? line length, angle and hopefully pull are known, so you should be able to gather most needed information by computing only reliable values. (try to fly a rev without your eyes, the only problem i know is to get propperly out of a fast spin; i guess you heard about controlling spin speed and how it is achieved!) : Any idea? hope mine do not go against how you presented your project... -grg- From - Thu Apr 18 08:15:44 1996 Path: gina.zfn.uni-bremen.de!informatik.uni-bremen.de!cs.tu-berlin.de!uni-erlangen.de!kue!lrz-muenchen.de!informatik.tu-muenchen.de!Germany.EU.net!howland.reston.ans.net!news-e2a.gnn.com!newstf01.news.aol.com!newsbf02.news.aol.com!not-for-mail From: birdofplay@aol.com (BirdofPlay) Newsgroups: rec.kites Subject: Re: AutoRev. Was Re: Laser-guided kite Date: 17 Apr 1996 13:12:29 -0400 Organization: America Online, Inc. (1-800-827-6364) Lines: 49 Sender: root@newsbf02.news.aol.com Message-ID: <4l38pt$8uq@newsbf02.news.aol.com> References: <4l0gfk$6ce@khyber.dev.madge.com> NNTP-Posting-Host: newsbf02.mail.aol.com X-Newsreader: AOL Offline Reader There you go .... A new Subject line and all !!! Great ! First Jeorg, your mail keeps bouncing back to me - got any ideas ? It happened to another who mailed me and we went to P Peters' site for another address of yours. Did you get any of my/our mail ? Second. I like Martin Gibson's ideas. The angular position was primitively floated earlier. I'm glad another corroberation is at hand ! I miss posted a suggestion to insert transducers into the kite lines of a human flier and take data and video in sync to get a feel for the ... Timing necessary, the power/pull needed to manage kite flight and for later use to "Teach" the robot. The notion of teaching the robot arms is, I believe, already known. Martins fly by wire is very intriguing. It's only one step from full autonomy. When a human can fly the kite via two joysticks then you will have made a serious stride in the overall task. I had earlier thought that a simple but fast winding mechanism would be perferable. Payout and control both on one mechanism. However, after considering the more powerfull industrial robot arms and their "reach" I have decided that this might be the better "first try" approach. Line extension could be easily added later. I'd go that way if I had the funding. BTW: You sly fox ! How'd you talk them into $30,000. ;-} Maybe you should be selling used cars ! just kidding ! You realize, of course, we are going to want to see some pictures of this gizmo in action ! If you choose ... You shoot them - I'll scan them and post them ! Keep us "in the know " Bob Neitzke birdofplay@aol.com http://members.aol.com/birdofplay From - Thu Apr 18 08:15:56 1996 Newsgroups: rec.kites Path: gina.zfn.uni-bremen.de!informatik.uni-bremen.de!cs.tu-berlin.de!uni-erlangen.de!kue!lrz-muenchen.de!informatik.tu-muenchen.de!Germany.EU.net!howland.reston.ans.net!tank.news.pipex.net!pipex!dish.news.pipex.net!pipex!laura.pcug.co.uk!tug!andrew From: andrew@tug.com (Andrew Beattie) Subject: Re: AutoRev. Was Re: Laser-guided kite Organization: /usr/lib/news/organisation Date: Wed, 17 Apr 1996 20:25:05 GMT Message-ID: References: <4kkjha$76p@newsbf02.news.aol.com> <31739FEC.2D2C@fbm.hs-bremen.de> <4l27vi$8ss@khyber.dev.madge.com> Lines: 31 mgibson@madge.com writes: >andrew@tug.com (Andrew Beattie) wrote: >>And build a feedback loop, so that the motor pulls in line to increase >>the pressure and lets it out to decrease it. > >I presume the intention here is to allow 3D flying, rather than always >relying on enough wind to fly the kite? I have a subtle distinction to explain here, on a subject that I have little understanding of (the more I study steering, the less I understand). I'll try my best. The whole thing that makes the kite steer is primarily the shift of the centre of tension between the bridle points. Ok, so we do this by moving the ends of the lines back and forth, but this is not the point. If we are measuring and manipulating the tension, I think that we have a much better control model. At any given time, we can measure the total tension in the lines, and adjust the control to deliver say 75/25 distribution of that tension between a given pair of lines. In this way the system will have a much better chance of coping with the variation in apparant windspeed that it will find not only from day to day, but also in different parts of the window. However, as you guess, such a generic approach should work in an extensive range of wind, from very strong, down through zero to low negative, just as human fliers do. (negative wind - ever done a 360 in other than dead calm?) Andrew -- # dirty hack to speed up compilation echo make | at now -30 minutes From - Thu Apr 18 07:30:16 1996 Message-ID: <3175D368.4E8B@fbm.hs-bremen.de> Date: Thu, 18 Apr 1996 07:30:16 +0200 From: "Jörg J. Buchholz" Organization: Hochschule Bremen X-Mailer: Mozilla 2.01 (Win95; I) MIME-Version: 1.0 Newsgroups: rec.kites Subject: Re: AutoRev. Was Re: Laser-guided kite References: <4kkjha$76p@newsbf02.news.aol.com> <4koj08$e8f@newsbf02.news.aol.com> <31739FEC.2D2C@fbm.hs-bremen.de> <4l2ilm$ct0@oops.ip.lu> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mozilla-Status: 0011 Content-Length: 957 blabla (is this your name?) wrote: > ... > as a human, you fly the rev by feeling (i do at least..), i.e. > you do not try a manouevre which the kite is not prepared to > follow, or by experience or by 'feeling'. > To integrate these parameters into a model is IMHO only possible > using fuzzy logic. i guess it will also be faster in calculus than > any algorithmic programming. > ... I have done some research on fuzzy control when I was designing flight controllers at DLR (our German NASA equivalent). IMHO they are a great Japanese advertising idea for selling "almost human" washing machines. If you have a good mathematical model using differential equations it is quite useless to try to translate this mathematical plant description into linguistic rules. RevSim has a good mathematical model of a Revolution kite. Thank you very much for your (other) ideas ;-) Joerg -- buchholz@fbm.hs-bremen.de (Jörg J. Buchholz) From - Thu Apr 18 08:12:58 1996 Message-ID: <3175DD6A.2974@fbm.hs-bremen.de> Date: Thu, 18 Apr 1996 08:12:58 +0200 From: "Jörg J. Buchholz" Organization: Hochschule Bremen X-Mailer: Mozilla 2.01 (Win95; I) MIME-Version: 1.0 Newsgroups: rec.kites Subject: Re: AutoRev. Was Re: Laser-guided kite References: <4l0gfk$6ce@khyber.dev.madge.com> <4l38pt$8uq@newsbf02.news.aol.com> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mozilla-Status: 0011 Content-Length: 1609 BirdofPlay wrote: > ... > Jeorg, your mail keeps bouncing back to me - got any ideas ? > It happened to another who mailed me and we went to P Peters' > site for another address of yours. Did you get any of my/our mail ? The guys at our computer center tell me they have "line problems". Anybody willing to untangle? > Martins fly by wire is very intriguing. It's only one step from full > autonomy. > When a human can fly the kite via two joysticks then you will have made > a serious stride in the overall task. I am sure this remote control via fly-by-wire will be an important intermediate step of the development process. Did you know that our young Dutch friend Gijs van Oort has already done the command input part for two-liners. He uses real-world handles and lines for the input of commands into his simulator. Gijs are you there? > I had earlier thought that a simple but fast winding mechanism would be > perferable. Payout and control both on one mechanism. However, after > considering the more powerfull industrial robot arms and their "reach" > I have decided that this might be the better "first try" approach. > Line extension could be easily added later. I am a fan of the KISS principle (Keep It Small and Simple). > I'd go that way if I had the funding. > BTW: > You sly fox ! How'd you talk them into $30,000. ;-} > Maybe you should be selling used cars ! just kidding ! Talk about 10% ($3,000) for the beginning. How much for a used robot with two arms? Legs required? Joerg -- buchholz@fbm.hs-bremen.de (Jörg J. Buchholz) From - Thu Apr 18 08:15:06 1996 Path: gina.zfn.uni-bremen.de!informatik.uni-bremen.de!cs.tu-berlin.de!zrz.TU-Berlin.DE!fu-berlin.de!golden-gate.owl.de!uni-paderborn.de!news.rwth-aachen.de!Aachen.Germany.EU.net!rmi.de!Germany.EU.net!Frankfurt.Germany.EU.net!news.maz.net!ins.net!blackbush.xlink.net!howland.reston.ans.net!newsfeed.internetmci.com!news.internetMCI.com!news-admin From: 4330291@mcimail.com Newsgroups: rec.kites Subject: Re: AutoRev. Date: 17 Apr 1996 11:11:31 GMT Organization: InternetMCI Lines: 18 Message-ID: <4l2jl3$5da@news.internetmci.com> References: <4kkjha$76p@newsbf02.news.aol.com> <4koj08$e8f@newsbf02.news.aol.com> <31739FEC.2D2C@fbm.hs-bremen.de> NNTP-Posting-Host: john.rcn.mci.com X-Newsreader: News X1.0-34 (Win32) Hopefully this thread will continue... I've been reading with some interest. Would the thought of Radio Control be applicable to the topic? Or is the idea for auto-control? Anyway, I've some experience with R/C, some with stunt kites, and none with kite building. Personally, I'd approach it with a single tethered line, with the servos operating control surfaces of some sort. I don't believe monster servos operating the lines(ie: from the ground as a person would) is the route to go... I'd be interested in any thoughts about this, or even if it's old hat and proved unfeasable. I'll be away about 10 days or so... Be awhile before I can pick back up on this. We shall see, eh? Stay High! -jrt From - Thu Apr 18 11:54:54 1996 Received: from emout13.mx.aol.com by hermes.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA02561; Thu, 18 Apr 1996 04:44:21 +0200 Received: by emout13.mail.aol.com (8.6.12/8.6.12) id WAA15360 for buchholz@fbm.hs-bremen.de; Wed, 17 Apr 1996 22:56:33 -0400 Date: Wed, 17 Apr 1996 22:56:33 -0400 From: BirdofPlay@aol.com Message-Id: <960417225631_516358080@emout13.mail.aol.com> To: buchholz@fbm.hs-bremen.de Subject: Re: [Fwd: Re: Laser-guided kite. Was Re: Bizarre kite laws] X-Mozilla-Status: 0011 Content-Length: 1545 Jeorg, I'm glad to see many more interested parties getting involved. I was also happy to see my ideas of angular position crop up again. Your point is well made re the tension monitoring as well as video position monitoring. It's the way humans do it and , in the computer, it's a cross check. Add wind speed to the data stream and you'll be able to predict/decide whether to attempt a maneuver or to wait for more or less air. With all this info comming in the clock speed are going to get bumped up or you'll have to use parallel processors. A MUXed A/D for each type of input might be a nice touch. One for line tension, another for angular position. I suppose that the A/D speed and required update speed will determine the is that is really necessary. But back to my original idea using transducers. It would be pretty simple to rig up some data collection on a laptop just to get an idea of the range of forces for a given wind speed. You could extrapolate up or down from whatever wind was blowing on the test day. With semiconductor straingages available these days , it should be a snap to lay down some bridges and gather a fair data base. I used to do gages when at a vehicle test track. 20yrs ago. Diff amps are a lot easier to keep balanced now as opposed to back in the old days. But then most of "electronics" is cookbook stuff these days. Well I don't want to take more time away from your fun Bye for now. Bob Neitzke birdofplay@aol.com http://members.aol.com From - Thu Apr 18 20:06:50 1996 Path: gina.zfn.uni-bremen.de!informatik.uni-bremen.de!cs.tu-berlin.de!zrz.TU-Berlin.DE!fu-berlin.de!news.belwue.de!news.uni-stuttgart.de!news.ruhr-uni-bochum.de!news.rhrz.uni-bonn.de!news.rwth-aachen.de!Aachen.Germany.EU.net!rmi.de!Germany.EU.net!EU.net!newsfeed.internetmci.com!swrinde!tank.news.pipex.net!pipex!lade.news.pipex.net!pipex!madge.co.uk!news From: mgibson@madge.com (Martin Gibson) Newsgroups: rec.kites Subject: Re: AutoRev. Was Re: Laser-guided kite Date: Wed, 17 Apr 1996 07:48:28 GMT Organization: Madge Networks Lines: 15 Message-ID: <4l27vi$8ss@khyber.dev.madge.com> References: <4kkjha$76p@newsbf02.news.aol.com> <4koj08$e8f@newsbf02.news.aol.com> <31739FEC.2D2C@fbm.hs-bremen.de> Reply-To: mgibson@madge.com NNTP-Posting-Host: 194.32.220.165 X-Newsreader: Forte Free Agent 1.0.82 andrew@tug.com (Andrew Beattie) wrote: >And build a feedback loop, so that the motor pulls in line to increase >the pressure and lets it out to decrease it. I presume the intention here is to allow 3D flying, rather than always relying on enough wind to fly the kite? I believe this is a feasible solution, but adds complexity too early on. When you can keep the kite in the air, then worry about refinements. Little, manageable steps are much more likely to result in success. Martin Gibson Madge Networks Ltd mgibson@madge.com From - Mon Apr 22 07:04:00 1996 Received: from relay1.pipex.net by hermes.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA11366; Sat, 20 Apr 1996 05:59:22 +0200 Received: from irene.pcug.co.uk by flow.pipex.net with SMTP (PP); Sat, 20 Apr 1996 05:11:35 +0100 Received: from Helen.PCUG.CO.UK by irene.pcug.co.uk id aa10700; 20 Apr 96 5:11 BST Received: from tug by helen.pcug.co.uk id aa14288; 20 Apr 96 5:11 BST From: Andrew Beattie Date: Fri, 19 Apr 1996 23:01:34 +0100 Newsgroups: rec.kites In-Reply-To: <317539B9.1FC7@fbm.hs-bremen.de> References: <4kkjha$76p@newsbf02.news.aol.com> <4koj08$e8f@newsbf02.news.aol.com> <31739FEC.2D2C@fbm.hs-bremen.de> Organization: /usr/lib/news/organisation X-Mailer: Mail User's Shell (7.2.4 2/2/92) To: buchholz@fbm.hs-bremen.de Subject: Re: AutoRev Message-Id: <9604192301.aa12164@tug.com> Status: U X-Mozilla-Status: 0011 Content-Length: 514 you write: >> And build a feedback loop, so that the motor pulls in line to increase >> the pressure and lets it out to decrease it. >You are definitely right. Thankyou. >On the other hand >you still need the outer visual feedback loop because it is not the >line tension you want to control but the position and the attitude >of the kite. Absolutely. I was only addressing one part of the problem... Andrew -- Windblade use mass unsolicited Email for marketing. Reject their product. From - Mon Apr 22 07:04:00 1996 Received: from pip.dknet.dk by hermes.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA11270; Sun, 21 Apr 1996 19:55:40 +0200 Received: from cph6.pip.dknet.dk (cph6.pip.dknet.dk [194.192.0.38]) by pip.dknet.dk (8.6.12/8.6.12) with SMTP id UAA02317 for ; Sun, 21 Apr 1996 20:07:58 +0200 Message-Id: <199604211807.UAA02317@pip.dknet.dk> From: rvo@pip.dknet.dk (Rolf V. Oestergaard) To: "Jörg J. Buchholz" Subject: Re: AutoRev. Was Re: Laser-guided kite Date: Sun, 21 Apr 1996 16:47:34 GMT X-Newsreader: Forte Free Agent 1.0.82 Status: U X-Mozilla-Status: 0011 Content-Length: 2382 On Tue, 16 Apr 1996 15:26:04 +0200, you wrote: >o What is the mean force per line of a Rev II? > What is the maximum force? Never broke a 25 kg lineset. But never fly in 10 m/s with it either, unless I use a vented kite. >o What is the maximum handle rotation speed (deg/sec) a human > pilot has to use in order to control the kite? > What about the maximum handle rotation acceleration (deg/sec/sec)? > Is it important? Depends on the handle length. I guesstimate I turn the handle 90 degrees in 1/10 second. >o Hydraulic or electric motors? Electric brushless DC motors. >o Power supply (batteries?, ac/dc?)? I don't have a car. A car battery and a switchmode power supply. Makes any voltage You want with like 80% effiency. Easier to design these days, with all the new parts for this. >o Is a simple video camera (Hi8?) with the usual video-in-card > for a PC sufficient for the resolution of let's say 10 degrees > of rolling angle? At which line length? I think the video rate of 50 Hz is too slow to control a Rev II !!! View one frame at a time, and see how much the kite turns... >o What would be the best color of the kite for the pattern > recognition? (blue sky?, white clouds?) Think in terms of signal colours. Not that important I think. >o Is a standard PC (P90, programmed in C) fast enough to do > pattern recognition and model following control in real time? Now that's a difficult question. If You can come up with efficient algorithms, then I guess so. >Any idea? You need to measure the tension on the four lines. This will provide faster feedback in the control loop than the video camera. Also this will be needed to avoid run-away, winding too lose etc. Maybe the thing to control should be the relative force on each of the four lines relative to each other, and not simply their "length". Somewhere in Your model, these values must be present also. Try to modify the revsim to use this as input, and see if the stearing gets easier. I think it will. Good luck. See You on FAN0 (kites allowed in windsurf area and during the festival), ______________ Rolf V. Oestergaard, Copenhagen, Denmark \ /\ / mailto:rvo@pip.dknet.dk http://login.dknet.dk/~rolf/ \/ \/ From - Mon Apr 22 07:13:37 1996 Message-ID: <317B1581.1235@fbm.hs-bremen.de> Date: Mon, 22 Apr 1996 07:13:37 +0200 From: "Jörg J. Buchholz" Organization: Hochschule Bremen X-Mailer: Mozilla 2.01 (Win95; I) MIME-Version: 1.0 To: "Rolf V. Oestergaard" Subject: Re: AutoRev References: <199604211807.UAA02317@pip.dknet.dk> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mozilla-Status: 0011 Content-Length: 273 Dear Rolf Thank you very much for your very detailled answers to my question. They will definitely be a great help in the days to come ... Are you working a job in one of those fields? Thank you again Joerg -- buchholz@fbm.hs-bremen.de (Jörg J. Buchholz) From - Mon Apr 22 08:43:17 1996 Path: gina.zfn.uni-bremen.de!informatik.uni-bremen.de!cs.tu-berlin.de!zrz.TU-Berlin.DE!fu-berlin.de!news.belwue.de!news.uni-stuttgart.de!uni-regensburg.de!lrz-muenchen.de!informatik.tu-muenchen.de!Germany.EU.net!howland.reston.ans.net!news-e2a.gnn.com!newstf01.news.aol.com!newsbf02.news.aol.com!not-for-mail From: birdofplay@aol.com (BirdofPlay) Newsgroups: rec.kites Subject: Re: AutoRev. Was Re: Laser-guided kite Date: 20 Apr 1996 15:56:49 -0400 Organization: America Online, Inc. (1-800-827-6364) Lines: 40 Sender: root@newsbf02.news.aol.com Message-ID: <4lbfi1$1k4@newsbf02.news.aol.com> References: <31739FEC.2D2C@fbm.hs-bremen.de> NNTP-Posting-Host: newsbf02.mail.aol.com X-Newsreader: AOL Offline Reader Jeorg, If a Fly By Wire is already inoperation then you have it much easier due to this prior work. >re: Gijs van Oort has already done the command input >part for two-liners. He uses real-world handles and lines for the input >of commands into his simulator. Gijs are you there? If there is ever any info re the FBW machine I'd like to see it. The mechanical part is almost done. Just add some tension feedback mechanisms like Andrew suggested. The tough part is the vision interface. I use a WinVision from Quanta for video grabbing from my VCAM. The B+W version is only ~ $150. Check Computer Shopper Mag for add. Have you determined the video resolution and frame speed necessary to get the job done ? Sounds like you need to start soliciting some donations from robot users / local business. They might even want to be associated the whole project. Aside from the neato techy aspects, resource managment is going to be one of the more challenging facets of the project. Donations from EC, due to shipping costs and duty issues, will be your best bet. There's gotta be some hitech "junk yards" that would be fun to visit in Germany. IMHO This thread is not as long as the amkite thread so I'd leave it here for now. This is more interesting and socially valuable too. Bob Neitzke birdofplay@aol.com http://members.aol.com/birdofplay Stay outa da treez From - Tue Apr 23 17:38:31 1996 Received: from pip.dknet.dk by hermes.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA09420; Tue, 23 Apr 1996 17:02:46 +0200 Received: from cph16.pip.dknet.dk (cph16.pip.dknet.dk [194.192.0.48]) by pip.dknet.dk (8.6.12/8.6.12) with SMTP id QAA11544 for ; Tue, 23 Apr 1996 16:59:01 +0200 Message-Id: <199604231459.QAA11544@pip.dknet.dk> X-Sender: rvo@pip.dknet.dk X-Mailer: Windows Eudora Light Version 1.5.2 Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Date: Tue, 23 Apr 1996 15:38:36 -0100 To: "=?iso-8859-1?Q?=22J=F6rg_J._Buchholz=22?=" From: "Rolf V. Oestergaard" Subject: Re: AutoRev Status: U X-Mozilla-Status: 0011 Content-Length: 1710 Joerg, >Thank you very much for your very detailled answers to my question. >They will definitely be a great help in the days to come ... Hope so, keep me informed on Your progress. Did some thinking yesterday, while flying my Rev II... I still think that You should go for a small fast control loop that can balance the load on each line according to the following inputs:=20 1) percentage load on each of the four lines lines 2) average line length This can be implemented as a sub-system of the program, or in a dedicated microprocessor, which also controls the motor phases directly (might save a couple of special cotrollers). The next step is a video to control-input loop. My observation was, that the fast stabilisation was done by feeling the load changing. I could keep the kite relatively stable, with my eyes closed. I think the video camera won't provide fast enough input to stabilize the kite. The delay will make it unstable. Try the fly-by-wire idea, and You will see. =20 >Are you working a job in one of those fields? Yes and no. I have a M.Sc.E.E, and work on my 8th year designing mainly digital electronics and software for satellite communication equipment, but also analouge circuits and small switching power supplies now and then. Servo systems, PID-regulators, digital regulators etc. I have from university, where these subjects were some of my favorites. -- ____________ Rolf V. =D8stergaard, Copenhagen, Denmark \ /\ / Home: rvo@pip.dknet.dk Work: rolf@login.dknet.dk \ / \ / URL: http://login.dknet.dk/~rolf/ \/ \/ From - Tue Apr 30 13:49:35 1996 Received: from shellx.best.com by hermes.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA11340; Tue, 30 Apr 1996 09:37:55 +0200 Received: from bsteele (bsteele.vip.best.com [205.149.161.160]) by shellx.best.com (8.6.12/8.6.5) with SMTP id WAA25890 for ; Mon, 29 Apr 1996 22:33:08 -0700 Message-Id: <3185A68F.78E8@best.com> Date: Mon, 29 Apr 1996 22:35:11 -0700 From: Bob Steele X-Mailer: Mozilla 2.0 (WinNT; I) Mime-Version: 1.0 To: "Jörg J. Buchholz" Subject: Re: Stunt Kite Simulator References: <3184506C.4B97@fbm.hs-bremen.de> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Status: U X-Mozilla-Status: 0011 Content-Length: 717 Gijs van Oort: sounds like you had a great time doing your simulator. try mine at: http://www.best.com/~bsteele/java/kite.html you'll need a java browser. i recommend netscape 2.0 for windows 95, nt or unix. you should be able to download it free from netscape. the simulation's physics model is not very good, i did it to learn java. i didn't spend much time on it but i've gotten some good reactions to it. i enjoy my XTC's more. see: http://www.best.com:80/~bsteele/kites.html happy flying, bob -- From - Wed May 22 06:57:17 1996 Received: from pip.dknet.dk by hermes.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA10505; Tue, 21 May 1996 21:53:46 +0200 Received: from rvo.pip.dknet.dk (cph2.pip.dknet.dk [194.192.0.34]) by pip.dknet.dk (8.6.12/8.6.12) with SMTP id WAA00244 for ; Tue, 21 May 1996 22:06:50 +0200 Message-Id: <1.5.4.32.19960521200645.00679934@pip.dknet.dk> X-Sender: rvo@pip.dknet.dk X-Mailer: Windows Eudora Light Version 1.5.4 (32) Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Date: Tue, 21 May 1996 22:06:45 +0200 To: "=?iso-8859-1?Q?=22J=F6rg_J._Buchholz=22?=" From: "Rolf V. Oestergaard" Subject: Re: AutoRev Status: U X-Mozilla-Status: 0011 Content-Length: 623 Hi, How's the autorev project going? Settled for a block-diagram of the control loops? Would love to see it! Another question: Can't run the RevSim at the moment, but wonder: Can You side-slide it? How? Mouse up/down is speed/brake and mouse left/right is rotation - how do You control a side-slide? Have fun, -- ____________ Rolf V. =D8stergaard, Copenhagen, Denmark \ /\ / Home: mailto:rvo@pip.dknet.dk http://login.dknet.dk/~rolf/ \ / \ / Work: mailto:rolf@login.dknet.dk (ro@tt.dk) \/ \/ From - Wed May 22 07:44:47 1996 Message-ID: <31A2A9CF.6E02@fbm.hs-bremen.de> Date: Wed, 22 May 1996 07:44:47 +0200 From: "Jörg J. Buchholz" Organization: Hochschule Bremen X-Mailer: Mozilla 2.02 (Win95; I) MIME-Version: 1.0 To: "Rolf V. Oestergaard" Subject: Re: AutoRev References: <1.5.4.32.19960521200645.00679934@pip.dknet.dk> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mozilla-Status: 0011 Content-Length: 2902 Dear Rolf > How's the autorev project going? Settled for a block-diagram of the control > loops? Would love to see it! AutoRev is prospering very well. Students and I have been doing some research on the line forces, handle speed and accelaration in different wind situations (necessary for choosing the right motors) and we have taken some first videos to train the pattern recognition algorithms. As you see, we are trying to approach AutoRev from the actuator and sensor sides. Its heart or brain, the control system, will not be developed until the rest of the body (eyes and arms) is healthy and working. :-( I will announce a www-page on the stepwise progress of AutoRev in rec.kites asap. > Another question: Can't run the RevSim at the moment, but wonder: Can You > side-slide it? How? Mouse up/down is speed/brake and mouse left/right is > rotation - how do You control a side-slide? rtfm ;-) >>>>>>>>>>>>>>>>>>>> part of the RevSim READ.ME file <<<<<<<<<<<<<<<<<<<< Beginner ========= Switch on the stabilizer, slowly push the mouse away from you to make the kite rise and try to stabilize it in the middle of the screen. Press both mouse buttons every now and then to equalize both handles and stop the kite from rolling. Alternatively you can use left and right mouse motions to compensate kite rolling. Whenever anything goes wrong, press the space bar to restart the simulation from a stable ground position. Avoid handle angles of more than 45 deg, because the kite behaves "strange". If the kites's reactions are too fast for you, double the damping parameters (elevation damping, azimuth damping, roll damping) in the REVSIM.DAT parameterfile. Advanced ======== Get a feeling on how the lateral control changes if the handle angles become greater than 45 deg. Let the kite roll on the ground until it is on its back and let it rise by pulling the mouse towards you. Find out that this only works with the stabilizer engaged. Try >>>>> sliding <<<<< the kite on the ground by pressing the >>>>> left <<<<< mouse button and moving the mouse sidewards. Experiment with different parameter in REVSIM.DAT (wind speed, kite mass, ...) Pro === Roll the kite 90 deg clockwise on the ground. Press the left button and move the mouse to the left, simulating a pulling of the left handle. The kite rises. Release the button and control the lateral motion by slowly pushing and pulling the mouse. Do the same maneuver in the air. Try it without stabilizer. Try to fly the kite on its back without stabilizer. You now need the left and right mouse buttons for manual stabilization. It's very hard. Try whatever you can do with your hardware revolution ... > Have fun, I will Joerg -- buchholz@fbm.hs-bremen.de (Jörg J. Buchholz) From - Fri May 24 07:33:50 1996 Received: from pip.dknet.dk by hermes.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA12218; Thu, 23 May 1996 19:38:13 +0200 Received: from rvo.pip.dknet.dk (cph27.pip.dknet.dk [194.192.0.59]) by pip.dknet.dk (8.6.12/8.6.12) with SMTP id TAA10507 for ; Thu, 23 May 1996 19:51:33 +0200 Message-Id: <1.5.4.32.19960523175124.0067f0a4@pip.dknet.dk> X-Sender: rvo@pip.dknet.dk X-Mailer: Windows Eudora Light Version 1.5.4 (32) Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Date: Thu, 23 May 1996 19:51:24 +0200 To: "=?iso-8859-1?Q?=22J=F6rg_J._Buchholz=22?=" From: "Rolf V. Oestergaard" Subject: Re: AutoRev X-Mozilla-Status: 0011 Content-Length: 657 Hi, >I will announce a www-page on the stepwise progress of AutoRev in rec.kites >asap. Great! I look forward to see something about the sensors - Consider using the motors to sense the force on the lines... >>>>>>>>>>>>>>>>>>>>> part of the RevSim READ.ME file =20 Thank You - I did not remember the use of the mouse buttons. Have fun, -- ____________ Rolf V. =D8stergaard, Copenhagen, Denmark \ /\ / Home: mailto:rvo@pip.dknet.dk http://login.dknet.dk/~rolf/ \ / \ / Work: mailto:rolf@login.dknet.dk (ro@tt.dk) \/ \/ From - Tue May 28 06:51:45 1996 Received: from pip.dknet.dk by hermes.rz.hs-bremen.de (AIX 3.2/UCB 5.64/4.280895) id AA14354; Sun, 26 May 1996 08:26:58 +0200 Received: from rvo.pip.dknet.dk (cph21.pip.dknet.dk [194.192.0.53]) by pip.dknet.dk (8.6.12/8.6.12) with SMTP id IAA11265 for ; Sun, 26 May 1996 08:40:26 +0200 Message-Id: <1.5.4.32.19960526063953.00667af0@pip.dknet.dk> X-Sender: rvo@pip.dknet.dk X-Mailer: Windows Eudora Light Version 1.5.4 (32) Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Date: Sun, 26 May 1996 08:39:53 +0200 To: "=?iso-8859-1?Q?=22J=F6rg_J._Buchholz=22?=" From: "Rolf V. Oestergaard" Subject: Re: line forces Status: U X-Mozilla-Status: 0011 Content-Length: 990 J=F6rg, >I've been thinking about measuring the motor current and take it as=20 >an indication for the forces and moments but then - I thought about=20 >using stepper motors, cause they give you better position accuracy. >And I fear its a bit more difficult to find a relation between >current and forces with the steppers. I don't think so - it's beeing done all the time. Consider using the stepper motors in the micro-stepper/brushless DC-motor configuration. Also You need feed-back from the motors - like an opto-coupled code-wheel or something - as You can't predict the exact nature of the controlled object - guess I don't need to tell You about that...;-) Have fun, -- ____________ Rolf V. =D8stergaard, Copenhagen, Denmark \ /\ / Home: mailto:rvo@pip.dknet.dk http://login.dknet.dk/~rolf/ \ / \ / Work: mailto:rolf@login.dknet.dk (ro@tt.dk) \/ \/ From - Fri May 24 07:58:40 1996 Message-ID: <31A55010.2EB0@fbm.hs-bremen.de> Date: Fri, 24 May 1996 07:58:40 +0200 From: "Jörg J. Buchholz" Organization: Hochschule Bremen X-Mailer: Mozilla 2.02 (Win95; I) MIME-Version: 1.0 To: "Rolf V. Oestergaard" Subject: line forces References: <1.5.4.32.19960523175124.0067f0a4@pip.dknet.dk> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mozilla-Status: 0001 Content-Length: 546 Rolf V. Oestergaard wrote: > Great! I look forward to see something about the sensors - Consider using > the motors to sense the force on the lines... I've been thinking about measuring the motor current and take it as an indication for the forces and moments but then - I thought about using stepper motors, cause they give you better position accuracy. And I fear its a bit more difficult to find a relation between current and forces with the steppers. We'll see ... Jörg -- buchholz@fbm.hs-bremen.de (Jörg J. Buchholz) From - Wed Jun 26 07:08:06 1996 Newsgroups: rec.kites Path: gina.zfn.uni-bremen.de!informatik.uni-bremen.de!cs.tu-berlin.de!uni-erlangen.de!uni-regensburg.de!lrz-muenchen.de!informatik.tu-muenchen.de!Germany.EU.net!howland.reston.ans.net!swrinde!newsfeed.internetmci.com!uunet!inXS.uu.net!psinntp!psinntp!psinntp!psinntp!se01!usenet From: neitzke Subject: Re: FLYING BLIND X-Nntp-Posting-Host: neitzke Content-Type: text/plain; charset=us-ascii Message-ID: <31CFFD68.172F@se01.elk.miles.com> Sender: usenet@se01.elk.miles.com (Usenet News) Content-Transfer-Encoding: 7bit Organization: Bayer Corporation References: <199606250727.DAA11492@endor.harvard.edu> Mime-Version: 1.0 Date: Tue, 25 Jun 1996 14:53:28 GMT X-Mailer: Mozilla 2.0 (WinNT; I) Back in 1984 I ran into a fellow in LA Calif. who flew Radio Control Models "by ear" !!! He too was blind. He was part of the Sepulveda basin RC buch. All of these stories and ideas add to the credibility/potential of computerized kite flying, being investigated currently in Gremany. Flying by "feel" ( pull and angle ) should be easily accomplished via computer control. Forget the video recognition algorythems. Human kite fliers don't NEED vision to get the job done. -- Bob Neitzke birdofplay@aol.com http://members.aol.com/birdofplay