jaco_robot model broken

[mattcorsaro1]

One of the fingers of the jaco_robot appears to be broken. https://imgur.com/a/dgr3r Nothing happens after I send an open command over graspit_interface. I have a fresh install of the latest version of graspit. Here's the terminal output when I import the hand:

Creating base...

Setting up 3 degrees of freedom...
Creating 3 kinematic chains (fingers)...
Chain 0:   Creating joints
   Joint 0
  Creating links
   Link 0
  Creating dynamic joints
Chain 1:   Creating joints
   Joint 0
  Creating links
   Link 0
  Creating dynamic joints
Chain 2:   Creating joints
   Joint 0
  Creating links
   Link 0
  Creating dynamic joints
No normalization data found; using factors of 1.0
Read 1 eigengrasps from EG file
Using eigengrasps from file: eigen/eigen.xml
Loaded virtual contacts from file virtual/contacts.xml

[jvarley]

Make this: https://github.com/graspit-simulator/graspit/blob/master/models/robots/jaco_robot/jaco_robot.xml#L50 d1+210 ----> d1+0 That will get the hand to show up correctly.

That hand does not play well with the dynamics engine as is. I am not sure why. In the jaco_robot.xml file try playing around with different values for

<maxEffort>-1.5e+5</maxEffort>
<Kd>1e+07</Kd>

maybe look at what has been used for the Barrett hand. Perhaps the weights, cog, or inertial matrices for the individual links are not correct.

https://github.com/graspit-simulator/graspit/blob/master/models/robots/jaco_robot/iv/6_hand_limb.xml#L6

If you get it working well, please make a pull request.

[JenniferBuehler]

After some updates to the graspit code, now there is a new version of the Jaco hand. The model has been generated with urdf2graspit from the jaco-arm-pkgs before it was added to this repo. This has been a while ago, so it could probably use an update. I will try to put a PR together soon. Meanwhile you could try to use the jaco_graspit_sample model here, does it work with that one?