uts-magic-lab / Magiks

Manipulator General Inverse Kinematic Solver
http://uts-magic-lab.github.io/Magiks/
Other
4 stars 1 forks source link

Using MAGIKS for other projects #2

Open FrecherFrechdax opened 5 years ago

FrecherFrechdax commented 5 years ago

Hello Nima,

We are currently working on an university project regarding the inverse kinematics of an planar manipulator. The number of DOFs is somwhere between 3-6 and should be flexible. The second priority tasks (collision avoidance, not moving all the joints with every movement, etc.) are very important. The trajectory is predefinde, we therefore have an trajectory projection scenario. The MAGIKS Toolbox looks perfect for this job! For two weeks we were trying to understand your code to implement this, but unfortunately didnt manage to achieve more than what already is in the Simple_IK script. Can you give us any idea for a start? Your help would be very apprechiated.

Thanks! Jonas and Johannes

nimasoft commented 5 years ago

Hi,

Thanks for your email. I am very happy that you are interested in my work. Unfortunately, I have changed my field and am currently working full time as a data scientist and haven't worked on MAGIKS for around 4 years and many things are forgotten. However, I will try my best to help you. Firstly, you need to define the geometry of your manipulator by DH parameters. Since they are planar, it should be easy as you can set all d parameters to zero. There are examples of geometries defined for some known robots like PR2, PA10 and PUMA. You can change that script to define your own geometry and make sure the forward kinematic engine gives the same result as you expect. Then you can specify an endeffector and define target and run the IK engine. You can also set a trajectory as target. First thing I recommend, is to run the example with one of the known geometries, (PA10 for example) to make sure it works for you. Then, change the geometry to your manipulator.

If you start doing that and had problems, I will be happy to answer any specific problems that you will face case by case, and we can go step by step. I am sorry, but unfortunately I did not have time to provide a comprehensive documentation for this product.

Best Regards Nima

On Thu, 11 Jul. 2019, 10:56 pm FrecherFrechdax, notifications@github.com wrote:

Hello Nima,

We are currently working on an university project regarding the inverse kinematics of an planar manipulator. The number of DOFs is somwhere between 3-6 and should be flexible. The second priority tasks (collision avoidance, not moving all the joints with every movement, etc.) are very important. The trajectory is predefinde, we therefore have an trajectory projection scenario. The MAGIKS Toolbox looks perfect for this job! For two weeks we were trying to understand your code to implement this, but unfortunately didnt manage to achieve more than what already is in the Simple_IK script. Can you give us any idea for a start? Your help would be very apprechiated.

Thanks! Jonas and Johannes

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/uts-magic-lab/Magiks/issues/2?email_source=notifications&email_token=AC4ZSPNHBTMFOQ4Q4XBENXLP64UYNA5CNFSM4IBEWDT2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4G6T66GA, or mute the thread https://github.com/notifications/unsubscribe-auth/AC4ZSPLANUZVLI2G2HY22ILP64UYNANCNFSM4IBEWDTQ .

awesomebytes commented 5 years ago

@FrecherFrechdax

Hello. Unrelated to this project, you could look into BioIK: https://github.com/TAMS-Group/bio_ik

It's a pretty nice project with software ready-to-use, a well written paper and some fancy videos to check out what it can do.

I've used it for different DOF (5-9) robot arms and it works.

I guess you want to use it from Python, so you could use: https://github.com/TAMS-Group/bio_ik_service

I used it to teleoperate the Arms of a Pepper from an OpenNI skeleton keeping the elbows as close as possible to the person pose in this snippet: https://gist.github.com/awesomebytes/745878504e4932f30de6e42e99193f27

Good luck!