search

CentroEPiaggio / pacman-DR53

This repository contains the necessary tools to run the demo contemplated in DR 5.3 of the PaCMan Project.
GNU General Public License v2.0
0 stars 0 forks source link

Grasp DB with IKEA objects #5

Open carlosjoserg opened 9 years ago

carlosjoserg commented 9 years ago

The current Grasp DB has too many errors in body tracking within a grasp sequence.

  1. After CentroEPiaggio/phase-space#2 is resolved and Object Pose Estimation is finished by @Tabjones, acquire grasps with IKEA objects to fill this dataset.
  2. Transfer acquisitions to the internal DB format (help wanted from @alessandrosettimi).
hamalMarino commented 9 years ago

At the current stage, there is no registered hand joint position: I would say that at least a few objects (first 5 as we agreed with @Tabjones) should have recorded grasp, then the rest will take some more time and be a matter of secondary importance. About transferring to internal DB format, this commit solves the issue: there is now a very simple utility (called grasp_storage) which allows this from command line with almost no effort (so, don't worry @alessandrosettimi !)

carlosjoserg commented 9 years ago

At the current stage, there is no registered hand joint position...

What do you need that for? If it is to discriminate the approach/grasp phases, the flexiforce readings can give you a good estimate for that as well, as long as you record them of course, which seems to be the case here

About transferring to internal DB format...

One question, the almost-no-effort is always that ctrl-c/ctrl-v of the timestamp from the bag play into the code, and then recompiling?

hamalMarino commented 9 years ago

Nope, have a look at the code if you want, but it's all command line... Still, you need the timing written somewhere, even in a file will work (try to rosrun dual_manipulation_grasp_db grasp_storage, I hope the help is clear enough!)

On Monday, April 13, 2015, Carlos J. Rosales notifications@github.com wrote:

At the current stage, there is no registered hand joint position...

What do you need that for? If it is to discriminate the approach/grasp phases, the flexiforce readings can give you a good estimate for that as well, as long as you record them of course, which seems to be the case here https://github.com/CentroEPiaggio/unipi-grasp-datasets/blob/master/pacman_wp2_grasp_db/data/startRecording.sh#L21

About transferring to internal DB format...

One question, the almost-no-effort is always that ctrl-c/ctrl-v of the timestamp from the bag play into the code, and then recompiling?

— Reply to this email directly or view it on GitHub https://github.com/CentroEPiaggio/pacman-DR53/issues/5#issuecomment-92347599 .

Ing. Hamal Marino PhD Student in Automation, Robotics, and Bioengineering

Research Center “E.Piaggio” Faculty of Engineering - University of Pisa Largo Lucio Lazzarino, 1 56122 Pisa - Italy Tel. +39.050.2217050 Fax +39.050.2217051

Email: hamal.marino@centropiaggio.unipi.it

http://www.linkedin.com/in/hamalmarino: linkedin.com/in/hamalmarino

carlosjoserg commented 9 years ago

Yes, it looks clear, thanks !