osrf-migration
commented
9 years ago Original comment by Bill Blank (Bitbucket: BillBlank).
I looked at the image. But the data looks good to me in both the v5 urdfs (plain and simple_shapes) and raw config files and doesn't seem to correlate with the image? Is it possible the engineer had attached 3rdparty hands to the robot?
Note that below both the moi and the mass of the hands are less than the two forearm pieces, and that in total the mass props appear fairly similar to the old hydraulic arms...
struct l_ufarm
index = 21
parent_link = l_larm
parent_kin_dof = l.arm.wry
mass = 2.12
com_x = 0.0
com_y = -0.2
com_z = 0.0
moi_xx = 0.006
moi_yy = 0.003
moi_zz = 0.006
moi_xy = 0.0
moi_xz = 0.0
moi_yz = 0.0
struct l_lfarm
index = 22
parent_link = l_ufarm
parent_kin_dof = l.arm.wrx
mass = 2.65
com_x = 0.0
com_y = -0.07
com_z = 0.0
moi_xx = 0.007
moi_yy = 0.004
moi_zz = 0.007
moi_xy = 0.0
moi_xz = 0.0
moi_yz = 0.0
struct l_hand
index = 23
parent_link = l_lfarm
parent_kin_dof = l.arm.wry2
mass = 0.53
com_x = 0.0
com_y = 0.06
com_z = 0.0
moi_xx = 0.00265
moi_yy = 0.00446
moi_zz = 0.00446
moi_xy = 0.0
moi_xz = 0.0
moi_yz = 0.0
Original report (archived issue) by Nate Koenig (Bitbucket: Nathan Koenig).
The original report had attachments: atlas_v4_simple_shapes.urdf, atlas_v5_simple_shapes.urdf
Atlas v5 might have incorrect hand inertias. See pull request #503.