Will likely build a small version to see .
Sadly some aspects of it (not the design itself, taking a closer look before I make up my mind) are just off:
- "open source" but you have to register to their mailing list first
- "open source" but they mention having "defensive patents"
- "open source" BUT CC non commercial...( I am always torn about this last one, so less important).
namely: a format for describing a hierarchy of parts , and constraints/joints :
- right now I have a tree like data structure something like
- but I also want to be able to define joints (such as prismatic, revolute, etc) , and their limits basically, something very similar to the URDF format (http://wiki.ros.org/urdf/XML/joint) .
The format i am working on is still intended to
- be simple
- json based
- and also needs to handle simple groups etc (which from what I gathered so far , are not handled all that well by the urdf format)
- multiple parents would be cool , but then again that potentially means more complex data structures (graphs)
- so my questions :
* does anybody who have worked with URDF know if it handles part trees as well ? (not clear from the documentation)
* are there any other similar efforts already underway that I have missed ?
* what are common issues when dealing with these kind of data structures ?
* any feedback at all is also welcome , thanks !
(If it is all confusing, blame my current cold)
The DHParameterKinematics class uses the Denavit-Hartenberg system https://www.youtube.com/watch?v=rA9tm0gTln8 to define links. the link ordering is crappy, its semantic to the XML (shame on me) in that the order of the links as they are defined in the xml is the order from base to tip of a chain of D-H links. I need to get around to adding a link order tag to that thing...
The link section is the mapping between the physical system and engineering units. It also specifies the limitations in the physical systems units, this is because the physical limitations should be absolute to the physical layer, since zeroing of the engineering units can mess with the limits, the limits are therefore implemented in the layer below.
The D-H parameters are properties of links, since links can be arranged in non-d-h definable systems. There are a few complex linkage kinematics implementations that do not use d-h definitions. The AbstractKinematics https://github.com/NeuronRobotics/java-bowler/blob/development/javasdk/NRSDK/src/com/neuronrobotics/sdk/addons/kinematics/AbstractKinematicsNR.java engine is one where the user can define arbitrary f-k and i-k implementations, and handles the wrapping of links in engineering units, the event system of conversions back and forth, interacting with the links, and even task-space path planning. see the enclosing directory for the link implementations.
I apologize for the lack of formal documentation on the AbstractKinematics system and D-H implementations, it is very new and a big project to migrate to open source quickly. I am willing to offer personal assistance implementing it with new users if you help with documenting as we go ;) (ill help anyway)
Documentation of the core Bowler Platform can be found here http://www.neuronrobotics.com/tutorials/ and some additional code examples on the new documentation site we are tryign to migrate over to: http://neuronrobotics.github.io/Java-Code-Library/Connection-Dialog-Example/
The hexapod was an early prototype and needs to be migrated over to the full LinkEngine system. It is more a rough draft of multi-kinematics-chain systems. It has a generated 5 down, one up smooth walk engine that will take any arbitrarily defined hexapod kinematics file. It uses the DyIO to coordination motion between all the servos for real time synchronization, and all kinematics takes place at the Java level.
What am I missing ?
- box corners printed
- box edges assembled
- drain system modeled ( a heavilly modified fork of 's Hydroponic drain)
- drain system printed and put into place (testing for now)
- YoumagineWeb App Engineer, 2014 - present
- InraBioInformatician/ Python Developper, 2011 - presentPython developper/ bio analyst
- Micro ApplicationQA/ C#.net Dev, 2006 - 2009QA Manager/C#/Nsis Developper
- Ciblo.netWeb dev, 2004 - 2004Web developper
- KITDev/3d printer specialist, 2014 - 2014
- IUT AmiensInformatics, 2003 - 2004
- CNAMInformatics, 2009 - 2011
- CNAMBiochemistry, 2009 - 2011
2013 review of developments, 3D Printing book release and more challenges.
#2014 I missed doing my annual roundup of 3D Printing adventures before Christmas, so here it is, with some plans and updates for 2014. Firs
Coffeescad: 3D-Druckmodellierung im Browser - Golem.de
Mit seinem Projekt Coffeescad will Mark Moissette nicht nur eine einfache Modellierungsanwendung für 3D-Drucker erstellen, sondern gleichzei
The Humble Indie Bundle #4 (pay what you want and help charity)
Pay what you want for a collection of awesome games, and help support two charities. All of the games are DRM-free and support Mac, Windows,
The electric revolution in R/C planes hits full-size aviation - DIY Drones
It's been amazing to watch how computer technologies have revolutionized the R/C aircraft world. The combination of high-efficiency brushles
Behold Tharsis Tholus, Mars' technicolor dream volcano!
You're looking at an image of Tharsis Tholus, an extinct Martian volcano that by Earthly standards is positively massive: towering 8km above
Free Template Alissa: Responsive Bootstrap Coming Soon Page | AZMIND - F...
Alissa: responsive coming soon template built with Twitter Bootstrap. It comes with images slideshow, countdown timer, AJAX email subscripti