Testing Newton for Industrial Robtics

Share with us how are you using the powerrrr of the force

Moderator: Alain

Postby DarthPhysics » Wed Jan 23, 2008 2:24 pm

Sure :D
Just finish the new Engine that we all can enjoy the real way of the Dark Side of the Force :twisted:
The theory is when we know all but nothing work.
The practice is when everything work but no one know why. <- that is me :) without the 'everything work'
User avatar
DarthPhysics
 
Posts: 44
Joined: Sun Apr 23, 2006 8:57 pm
Location: Boulay-Moselle , France

Postby DarthPhysics » Wed Jan 23, 2008 5:05 pm

The force is with me :D

look :

http://fr.youtube.com/watch?v=yD96HgPnY6Q

i use now the cadencer process from oxNewton :wink:
Alot better , for more better i have to kill the halve of my program and star again :shock:

Cya latter...
The theory is when we know all but nothing work.
The practice is when everything work but no one know why. <- that is me :) without the 'everything work'
User avatar
DarthPhysics
 
Posts: 44
Joined: Sun Apr 23, 2006 8:57 pm
Location: Boulay-Moselle , France

Postby Julio Jerez » Wed Jan 23, 2008 5:29 pm

As soon as I get son time, I will put some time to help you out to get a more stable response.
But it is very good that you are trying to get it by yourself :D

Man It is amazing that the whole thing works. :shock:
Are you getting the angles to control the real robot?
Julio Jerez
Moderator
Moderator
 
Posts: 12249
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Postby DarthPhysics » Wed Jan 23, 2008 6:49 pm

Thanx :wink:

I dont get the angles for each axis now.
I prefer first do a stable and fast movement + orientation with a maximum +/- 1mm Tolerance.
That tolerance will be enough for my Robot , but by the way i want really
simulate one industrial robot , the tolerance have to go at 0.1 to max 0.3mm.
It's alot ambush on a padawan path , i'm not ready to kill my robot now , i prefer do a strong and precise soft before i start with hardware communication :D
I can use the robot without the PC , that's the role from the 2 controller card you see on the photo.
For doing this i connect a usb keyboard to the controller and enter the command i want see the robot do.
But it's only Axis movement , Newton will help for move the TCP directly :wink:
You say :

As soon as I get son time, I will put some time to help you out to get a more stable response


Nah , dont get time for me before you finish the new version , i can wait :wink:

Cya :D
The theory is when we know all but nothing work.
The practice is when everything work but no one know why. <- that is me :) without the 'everything work'
User avatar
DarthPhysics
 
Posts: 44
Joined: Sun Apr 23, 2006 8:57 pm
Location: Boulay-Moselle , France

Postby DarthPhysics » Thu Jan 24, 2008 11:44 am

The Today News :D

After working arround for do better and smooth movement i have recode
some part of my path code in the cadencer.
And adding a reader for get a file with movement that the robot read and
execute at runtime.
Now all is very smooth and natural , like a 'real' robot.
It's just some 'jump' when the robot have to turn his axis too fast for
try to go in some tool orientation/position.
The video is here :
http://fr.youtube.com/watch?v=bf1x424Coss

Cya latter.
The theory is when we know all but nothing work.
The practice is when everything work but no one know why. <- that is me :) without the 'everything work'
User avatar
DarthPhysics
 
Posts: 44
Joined: Sun Apr 23, 2006 8:57 pm
Location: Boulay-Moselle , France

Postby DarthPhysics » Mon Feb 04, 2008 11:46 am

Some news today :D

Alot modification in my coding for control the robot end effector give me now
a very smooth movement and orientation.
A little problem with the TCP du to the non collision state when i position/orient directly the end effetor.
But the robot follow without jump etc :D
I have do a little video for show you the actual state.
http://fr.youtube.com/watch?v=ATlRt8dW9Fc

anyway , thanx to k00m for all his goood advice :wink:

Regards.
The theory is when we know all but nothing work.
The practice is when everything work but no one know why. <- that is me :) without the 'everything work'
User avatar
DarthPhysics
 
Posts: 44
Joined: Sun Apr 23, 2006 8:57 pm
Location: Boulay-Moselle , France

Postby DarthPhysics » Mon Feb 04, 2008 3:08 pm

Heya again :D

The robot whas a little ...slow ...
I have now add some change to my sequencer , now the TCP go lightspeed but the robot have some problem to follow at that speed :wink:
But i'm sure Master k00m or Julio will know how i can arrange this :D

video here

http://fr.youtube.com/watch?v=7R6hQYvwG5s

Cya :D
The theory is when we know all but nothing work.
The practice is when everything work but no one know why. <- that is me :) without the 'everything work'
User avatar
DarthPhysics
 
Posts: 44
Joined: Sun Apr 23, 2006 8:57 pm
Location: Boulay-Moselle , France

Postby Julio Jerez » Mon Feb 04, 2008 3:15 pm

what you do? In it not too bad

thi si sgioing fo rteh
Kickkkkkkkkkkk Asssssssssss award
Julio Jerez
Moderator
Moderator
 
Posts: 12249
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Postby Dave Gravel » Mon Feb 04, 2008 7:32 pm

About this new video:
In oxNewton I have add some function to change visual 3d position and rotation directly in runtime and it update the newton obj matrix.
Normaly I use this function to reset the object at the default start position in a scene.

In my idea it is not really the best solution to use it constantly,
because the time that object is updated it lose newton support.
It is surely why that TCP make a lag between the joints and the robot speed.
I have explain to Darth what is this function tricking and why it behav like this but currently he say it is the best result and the most stable method tested.

Sure in the futur he need to use a other solution or something like this but better interpreted.
Anyway I have seen the difference from the first test and yes it looking a lot more stable with this tricking.

Nice work Darth and good continuation, I very like to see this robot in action.

PS:
It is just to make sure Julio understand the problem and the method that you use currently.
You search a nice physics solution, if you can read this message you're at the good place :wink:
OrionX3D Projects & Demos:
https://orionx3d.sytes.net
https://www.facebook.com/dave.gravel1
https://www.youtube.com/user/EvadLevarg/videos
User avatar
Dave Gravel
 
Posts: 800
Joined: Sat Apr 01, 2006 9:31 pm
Location: Quebec in Canada.

Postby DarthPhysics » Mon Feb 04, 2008 9:06 pm

You are right master k00m :wink:
It's not ..THE .. solution for the future use but atm it's the best i have for move and orient precisely the TCP.
The best is probably to wait that the new version come out , then , if possible we can then look for implement anything that approach a real robot end effector.
But anyway , with NGD & oxNewton i get alot better results in the robot movement than every other system i have do before :wink:
And the 'other' best is that i get really high FPS with this ...
That will be suitable for the future version when i will simulate more than only 1 robot in real time :wink:

Anyway i'm happy that you like it :D

Cya.
The theory is when we know all but nothing work.
The practice is when everything work but no one know why. <- that is me :) without the 'everything work'
User avatar
DarthPhysics
 
Posts: 44
Joined: Sun Apr 23, 2006 8:57 pm
Location: Boulay-Moselle , France

Postby DarthPhysics » Tue Feb 05, 2008 12:39 pm

Hi :D

@Julio :

Is it possible to simulate a gear ?
When i try , it's like my gear is considered as a simple round object.
I know they is some problem with convex object in other physics engine , but what about NGD and convex ?

Cya :wink:
The theory is when we know all but nothing work.
The practice is when everything work but no one know why. <- that is me :) without the 'everything work'
User avatar
DarthPhysics
 
Posts: 44
Joined: Sun Apr 23, 2006 8:57 pm
Location: Boulay-Moselle , France

Postby DarthPhysics » Wed Feb 06, 2008 12:50 pm

Hello all :D

Just because is not fair to stay alone in a empty world , i added some
friends to my robot :wink:

just take a look here :
http://fr.youtube.com/watch?v=UnQSqrrAPyo

Cya :D
The theory is when we know all but nothing work.
The practice is when everything work but no one know why. <- that is me :) without the 'everything work'
User avatar
DarthPhysics
 
Posts: 44
Joined: Sun Apr 23, 2006 8:57 pm
Location: Boulay-Moselle , France

Postby VeTaL » Wed Feb 06, 2008 1:21 pm

its amazing ))))))
User avatar
VeTaL
 
Posts: 5
Joined: Sun Mar 06, 2005 4:05 pm

Postby Julio Jerez » Wed Feb 06, 2008 1:50 pm

Om mind the dancing robot. Wiow said robot can not dance the macharena?
Julio Jerez
Moderator
Moderator
 
Posts: 12249
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Postby DarthPhysics » Wed Feb 06, 2008 3:05 pm

Heya again :D

For 'macharena' i dont know but all is possible with robots :)
I have set different sequence for each robot in this video.
http://fr.youtube.com/watch?v=XSJPz1WAOyQ

As you can see , each robot is perfectly synchronized :D
Master k00m tell me it will be better and faster with the next version
of NGD/oxNewton , then .... wait and see :wink:

Cya and thanx for making this great physics engine :D
The theory is when we know all but nothing work.
The practice is when everything work but no one know why. <- that is me :) without the 'everything work'
User avatar
DarthPhysics
 
Posts: 44
Joined: Sun Apr 23, 2006 8:57 pm
Location: Boulay-Moselle , France

PreviousNext

Return to User Gallery

Who is online

Users browsing this forum: No registered users and 12 guests

cron