Model CantiBox price quote.

[yck011522]

From the rough setting out file Davide sent me 20211005_DT.3dm

[yck011522]

Even only for the sake of the I ran check state: p -m integral_timber_joints.planning.check_states --problem cantibox_process.json --design_dir 210813_ThreeBox --reinit_tool --batch_run Results:

#1155 A244M0 fail
#1204 A255M4 fail

Check IK: p -m integral_timber_joints.planning.check_ik --problem cantibox_process.json --design_dir 210813_ThreeBox --gantry_attempt 100

Results as follows:

Failed Movement id | State Indices | Error msg
('A3_M1', 13, 'no IK solotion found after 100 gantry attempts | total time 35.972')
('A8_M1', 33, 'no IK solotion found after 100 gantry attempts | total time 33.506')
('A42_M0', 213, 'no IK solotion found after 100 gantry attempts | total time 9.421')
('A48_M6', 247, 'no IK solotion found after 100 gantry attempts | total time 11.430')
('A244_M0', 1156, 'no IK solotion found after 100 gantry attempts | total time 27.374')
('A244_M1', 1157, 'no IK solotion found after 100 gantry attempts | total time 29.312')
Total checking time: 425.090

[yck011522]

13 and 33 is related to #126 213 and 247 is related to a clamp placement that can be flipped and fixed. 1156 and 1157 is related to the center vertical piece on the tilted fact crashing with the ground pre-assembly due to the tilt. This piece is easier if it can be shortened. If not we can fit in from inside but we will need to check if the robot fits.

[yck011522]

Fixed 3 things:

• Fixing incorrectly modeled env_models
• Flipping the clamp attaching pose for the vertical beam
• Flipping the middle beam assembly direction to come from inside.

p -m integral_timber_joints.planning.check_states --problem cantibox_left_process.json --design_dir 210813_ThreeBox --reinit_tool --batch_run`
p -m integral_timber_joints.planning.check_ik --problem cantibox_left_process.json --design_dir 210813_ThreeBox --gantry_attempt 100

check_state all passed. check IK fail for 13 and 33. Not sure why, still investigating.

[yck011522]

Turns out for 13 and 33, the gripper tip is just touching the ground block.

• Option 1 : Model the gound pillow block to be slightly smaller
• Option 2: Change Grasp to another location such that they don't exactly kiss the pillow block
• Option 3: Change the gripper jaw hardware and thus the collision model can be smaller. (Not ideal)

Probably will go with Option 1.

[yck011522]

This is closed by c293fdb408a22abb4370d14935e3cf91a2c6f328