AMOR - Automated Milking Ongoing Robot

I’m sorry for not answering you as you deserve it. Just watched „Das Geheimnis der Hebamme“ and it gave me a motivation boost. We have to make it happen. We have to make it happen for the „white sheep“ for the good-hearted people out there. I’m so sorry that I lost my belief and trust in this world. To idle around watching isn’t helpful, one has to take action even if one thinks this action is all void (just as it occurred to me so hopeless recently).

Let’s make it happen. We have to take action.

Forgive me please that I have no specific tasks.
I would hate myself giving orders or tasks to you, you are free, I am an advocate of free world and scientifically, critically discussed law based on common interest, public domain, objectivity.

If you wish then I’d be glad if you could register on http://github.org and set up a development environment. https://git-scm.com/book/en/v2/Getting-Started-About-Version-Control (People use Git for software mainly, we use it for hardware and everything else too unless we find and discuss arguments contra git for all or pro another tool. Github is used for hosting of e.g. git repositories to not overload our own servers, nice of Github to be free for public domain projects! Thanks!)

That would be very kind and cool to contribute to the repository together with you and all the others that are willing to help. We could define some early tasks like stepper motor controller and the overall controller. Each will get its own git repository and the AMOR overall repository collection will put all the repositories it needs together using git submodules, It may be too much at once to read up on submodules now, I don’t want to overload or exploit you.

AMOR news

  • The worm manipulator is on hold due to:
  • The mechanics (rope redirection without risk of breaking ropes, i.e. using miniature roll bearings) getting too complicate to be worth the benefit (which is the ball bearing functionality, thus reducing axis count and also the weight saving as the motors could be mounted at the base, not on the joints).
  • Each link joint rotation angle was limited to 45°, at maximum 60°.
  • The worm robotic manipulator may be catched up later again if we get a proper electro-magnetic technological base which we are currently lacking in open source technology.
  • The new goal is a robotic 6-axis robot. There’s already some progress. It’s a surprise for you all good hearted beings!

Don’t be sad, you are not alone.

There is a world development throng upcoming.

Would You be able and willing to join a mechanical plant-removal throng this year? Could be as early as tomorrow. To compensate I’ll help you with removing these plants from your uncle’s farm.
Be warned the plant has spread drastically due to compressed ground et alia on the meadows.

I think the calf milking robot variant not even is that hard. In a sense it’s even easier other than the crazy odometry - might drive along using auto-correction naturally or artificially (maybe a sensor where the milk chamber / container is located).

Much harder is the robotic manipulator (which I’m are currently sourcing parts for) or the 3D camera system or what else complex technology I forgot.

AMOR news

Failure to source spur gear in extra width, won’t order custom made because I want all AMOR parts to be easily available or easily custom-made and a gear isn’t easily custom-made for villages like the one out of the movie I cited in the post before :wink: - therefore switched joint transmission system from ultra-reduction ratio with low…medium torque due to mechanical slip mechanism to a double compound gear system. Wish me luck sourcing this compound spur gear … if impossible to source then several (3) gears will be welded - hopefully with proper tooth phase alignment.

AMOR news

Breakthrough: Door system: 4 exit directions, one cylinder only. Absolute simplicity. Drawings pending.

Combined with an interconnected 3 height ramp lever floor/roof system one of the 4 cow exits expands to 3, resulting in 6 exit directions using only 3 cylinders. Mighty cow and cloudy sky!

(Just found this solution while thinking about what to do with a remainder of tile glue. Started putting tiles onto the milking robot wall towards the crazily constructed milk chamber.) (Another government cross compliance check is due sometime the next days.)

https://www.landwirtschaftskammer.de/landwirtschaft/tiergesundheit/rgd/melkroboter.htm

Thanks, great resource.

AMOR Update

  • It was managed to hugely simplify the hardware and software, using only x86_64 architecture consistently.
  • Full Desktop system running Voidlinux + Xmonad serves as watchdog, main control and image processing network (no special code required for ARM BeagleBoneBlack or other Microcontrollers - no risk that the processing power of these systems doesn’t suffice for the huge amount of image data,no risk that LinuxEMC2 support is limited).
  • Because LinuxEMC2 is used for the new universal 6 axis manipulator inverse and direct kinematics, time is saved, reinventing the wheel is prevented.
  • The touchscreen is run by a hardened Gentoo + Gnome or Unity (which are the only Linux window managers supporting touch gestures somewhat properly).
  • The option to approach cows on the fields is added to the overview schematic.

https://github.com/faerietree/autonomous_milking_ongoing_robot/blob/master/AMOR_High_Level_Overview.jpg (view Raw to enlarge)
This is a major step forward, worlddevelopment e.V. will try to get the manipulator to work first. A multitude of our Global Village Construction Set’s modules depend on this development.

Autonomous Milking Online Robot Update
The next Computer Network Design Breakthrough
High Level Overview (as of commit b005d457781e555ebbeb2cd8d4ebd0974211ad5), increased resolution:
https://github.com/faerietree/autonomous_milking_ongoing_robot/raw/master/AMOR_High_Level_Overview.jpg

  • Using ARM Jetson + Ubuntu + ROS + OpenCV for Image processing. This has been proven to work epicly by others before.
    .
  • Centralize all image processing to the server (which most of the time has nothing to do anyway). This further reduces costs because no matter how many robotic arms are used, it’s all possible using one 3D stereoscopic camera + 1 Jetson Quad Core which is mere 190€ (compared to Desktop computers with equivalent power this is virtually nothing!).
    .
  • Specify the parallel port for the manipulator et alia controller, i.e. the void linux realtime system. Can be replaced by Arduino(s) if preferred, though note that use of LinuxEMC2 is not so easy then (may work on BeagleBoneBlack).

.
It seems to emerge that the cost will be far far less than the commercially available milking robots. We are talking about some few thousands total cost here.

You can find worlddevelopment Satzung Draft on Github - note it’s only available in German, because the club’s legal seat is located in Germany.

It is to be decided in the next member throng which stance worlddevelopment will take towards UNICEF and opensourceecology, i.e. which will take membership where. Note active membership in worlddevelopment is 1€ a year to be fair. Fördermitgliedschaft / passive / donor membership can be anything from 1 € up.

I remember my Transall navigator uncle tell me from his experience with UNICEF et alia. This is a longer story thus best not told here in the AMOR forum.

AMOR news:

Because AMOR is entirely open source and modular to be built from scratch, it features its own low tech transmission design.
This joint transmission was reworked to allow lifting a payload of 4 Kg in the worst case scenario (longest lever, i.e. straight manipulator arm).

You can not find the design on Github because development is a bit ahead of documentation. It’s binary files too. Also it’s low-tech and will be tested once I’ve earned enough money to order the parts.

Worlddevelopment has been founded as a non registered club which turned out to be the most productive way. It suffices when OSEG gets the e.V.
We then could be member in OSEG. This is still to be discussed in more detail.

Worlddevelopment is collaborating closely with HAMAG Maschinenbau, Mittelmair Handels GmbH repair, construction facilities and energy harvest, blenderartists, buildtheenterprise and hackerspaces as well as individual freelancers.

One of the new projects created in collaboration with Mittelmair Handels GmbH is an open source industrial bending machine.

AMOR news:

Because we thought about how to end the lack of funding for AMOR and other projects, it was decided to rent out an old site. This undertaking will in the long run improve our funds and ensure our independence.
Currently the initial costs are requiring all funds we can somehow earn so there is still nothing left for AMOR. The funds will still be very tight for the next year. It is nothing new for me to nourish from oat and tap water + mashed apples which is the cheapest I can find. At the moment we are dependend on gifts from family members, e.g. on our wishlist there is the Jetson Tk1 to start code development.

So if you want to help bridge this development void you can (if you like) donate to radagast@dragontale.de via PayPal but we can not give a receipt.

Vielleicht etwas off-topic, aber ich bin grade auf eine coole Sache namens „WebGL“ gestossen und in diesem Zusammenhang auf ein Blender-Modul namens „Blend4Web“. Hier ein Beispiel einer Molkerei, welches mit Blend4Web realisiert wurde:

https://www.blend4web.com/en/demo/dairy_plant/

Gruss, Oliver

Interessant koennte insbesondere sein, dass mehrere Molkerei-Produktionsschritte detailliert im Schema gezeigt werden. Vielleicht koennen wir so einfacher freie Kaese-Maschinen rekonstruieren.


Weil weiter oben bereits ueber die limitierten AMOR Finanzen gesprochen wurde:

Erster Einblick in Worlddevelopment Finanzen
Wolltest du schon immer wissen, wie eine Nebenkostenabrechnung auszusehen hat? Oder wie viel Müll, Kanal, Trinkwasser, Elektrizität, Kaminkehrer in Süddeutschland nahe Schloss Neuschwanstein kosten?

Dann wirf’ ein Auge auf die ersten veröffentlichten Umsätze unseres Worlddevelopment Vereins. http://facebook.com/foranepicworld/

ja, das mit der Molkerei war jetzt auch nur mal so ein beispiel was mir grad so vor die Flinte kam, aber ich dachte das Dich das vielleicht interessiert auch wegen „Blend4Web“.

Was man m.E. so richtig cooles damit machen kann ist z.B. dieses hier:
http://3dcontent.be/axis_8mm_rods/axis_8mm_rods.html

Ist wirklich nett fuer Dokumentation. Dann muss man keine Screenshots mehr machen sondern das 3D model kann gleich direkt vom Git repository gerendered werden. (WebGL)

Vll in Zukunft sogar gemeinschaftlich an ein und demselben 3D Modell arbeiten, aber mir fehlt dann die Versionierungskomponente. Ein Entwicklungsprozess ist so riskant, dass man einfach manchmal wieder auf einen alten Stand zurueckgehen koennen muss oder zumindest sehen muss, was man wann wie geaendert hat.

Interessant ist auch, wie wechselhaft die Leute sind, zuerst die Tendenz alles als WebApp zu haben (online games, … hohe Serverlast, geringerer User Stromverbrauch) … dann kommt die Smartphone Phase und alles wird wieder umgedreht, also wieder wie frueher bei DesktopApps die Rechenleistung wieder zurueck auf die Nutzergeraete.

Jetzt WebGL wieder anders herum. Hin und her, her und hin.

Trotzdem, fuer Doku ist das der Hammer, wenn nur jemand so einen Service programmieren koennte, der:

  • die Grafik/Datei von einer WebAdresse per GET parameter einliest. [1]
  • und mittels WebGL rendered mit solch nuetzlichen Controls wie „Explode“ wie zu sehen unter deinem Link
    Dann koennte man direkt vom Wiki oder von der README.md von Github aus darauf verlinken (deshalb der GET parameter).

Ich koennte das programmieren. Aber dafuer fehlen mir saemtliche Mittel. Als Open Source Ingenieur muss man sowieso immer wieder was programmieren weil irgendetwas nicht oder noch nicht geht. Ah … das ist doch alles ein Fluch.

[1] Dateiformate von Relevanz sind Blender|FreeCAD oder nur Wavefront.OBJ (STL kann keine Farben speichern). Am besten waere gar COLLADA weil es auch Animationen speichert, aber ob sich das wohl jemals als Standard durchsetzt, ich glaub nicht mehr daran, obwohl ich selbst es oft verwende und auch weiter verwenden werde.

Hab auch Neuigkeiten zu AMOR:

Noch immer keine einzige Spende (die Leute sind zu beschaeftigt um fuer so Kampagnen wie „Spende und wir verkaufen dir die Creme, die dich wieder aussehen laesst wie mit 20“ ihre Millionen zu investieren). Einstein hatte recht … und das macht es nicht unbedingt leichter, sich fuer die Gemeinschaft einzusetzen.
Hier in der Gegend treten jetzt wieder mehr Leute von ihren Ehrenaemtern zurueck, weil „das Ehrenamt nicht mehr angesehen ist“.

Tja. Wie auch immer worlddevelopment wird sich so einfach nicht aufhalten lassen (da muss man uns schon liquidieren).
Jedenfalls Walker macht Fortschritte.

Und AMOR erst recht, aber fuer Doku fehlt mir saemtliche Zeit. Eins wird klar sein, wenn der Gemuese Frucht Nuss Roboter rauskommt, dann werden die Leute staunen aber richtig freuen wuerd es mich erst, wenn es auch liebe Leute nachbauen koennten! (aber ohne Doku?)
Fuer sowas muss man so viele verschiedene Technologien beherrschen von Mechanik, Schweissen, Hydraulik ueber Elektrik, Elektronik, Loeten bis hin zu Hardware- und Web/Service Programmierung. (aber ohne Doku?)

Ich halt’s mittlerweile wie Jan Slinsky. Ich kaempf fuer die „Gute“ Seite der Macht im kleinen Kreis, da kann man sich ja so schon kaum durchsetzen, man muss immer streiten, wissenschaftliche Argumente glaubt dir doch eh keiner.

AMOR NEWS

  • 5 axis Manipulator Mechanik ist designed.
  • voll hydraulisch (elektrisch hat einfach nicht mehr gereicht mit den 10Kg payload - ansonsten wird die Transmission alles andere als DIY!!)
  • trajectory capable
  • scalable (use as crane, e.g. for clearing animal dung, forest work, removing waste food, …)
  • state of the art technology: void Linux, BBB, JetsonTk1

LG, der Oeko Typ (der die Welt fuer eine Mistwelt haelt, zumindest zeigt sie mir es jeden Tag lol sch* Welt :mrgreen: und viel zu wenige helfen der Freien Welt mit Open source Entwicklung oder Spenden … immerhin OSEG ist eine der Ausnahmen! Herzlichen Dank)

Changelog
https://github.com/faerietree/manipulator/commits/master

Fahrplan:

  1. Manipulator
  2. Spezialanwendung: melken|andocken

Roadmap:

  1. Manipulator
  2. special application: milk|dock

Cost estimation: 1000 Euro for the manipulator. 98% DIY, simple (no specially shaped parts). 2%: motors, super computer