INFO] optimization couldn't satisfy all constraints when roslaunch tesseract_ros_examples puzzle_piece_auxillary_axes_example.launch

[dbdxnuliba]

roslaunch tesseract_ros_examples puzzle_piece_auxillary_axes_example.launch

INFO] optimization couldn't satisfy all constraints Info: TRAJOPT motion planning failed (Failed to find valid solution) for process input: Tesseract Composite Instruction Debug: worker 4 finished running TRAJOPT Debug: worker 0 ready to run Error Callback

[Levi-Armstrong]

Are you running 0.13.X version?

[dbdxnuliba]

it's 0.12.0

[Levi-Armstrong]

Interesting. I will see if I can track down the release where it broke.

[Levi-Armstrong]

This was broken when we switched OSQP to enable adaptive_rho by default which in most cases speeds up convergences, but in this case it prevents it from converging. I am exposing the convex solver config in the trajopt solver profile so this can be disabled.

[Levi-Armstrong]

Fixed by recent PR.

[dbdxnuliba]

where can I modify the config param

This was broken when we switched OSQP to enable adaptive_rho by default which in most cases speeds up convergences, but in this case it prevents it from converging. I am exposing the convex solver config in the trajopt solver profile so this can be disabled.

[Levi-Armstrong]

where can I modify the config param

This was broken when we switched OSQP to enable adaptive_rho by default which in most cases speeds up convergences, but in this case it prevents it from converging. I am exposing the convex solver config in the trajopt solver profile so this can be disabled.

Here is what I had to modify

https://github.com/tesseract-robotics/tesseract_planning/blob/00726c502b3a71c2df02576e0012037884e4ef6a/tesseract_examples/src/puzzle_piece_auxillary_axes_example.cpp#L211-L218