Can I modify the parameters after loading udrf？

[lcb512]

I want to do experiments on exoskeleton robots with different leg lengths. Can I modify the parameters (size, quality) after loading udrf？

[jcarpent]

Yes, of course, you can modify your model after loading: change inertia, joint placements, etc.

[lcb512]

I want to modify the leg length in rqt instead of directly modifying it in urdf. I don't know what to do. Is there any good way?

[jcarpent]

Following the nice answer of @stephane-caron, in Python you can easily do:

model.jointPlacements[model.getJointId("my_joint_name")].translation = np.array([1.,2.,3.])

[lcb512]

Ok, I understand,what is the code for C++.

[jcarpent]

Exactly the same, except you should use Eigen:

model.jointPlacements[model.getJointId("my_joint_name")].translation << 1.,2.,3.;

[lcb512]

Thank you very much. I'll try it.

[stephane-caron]

Proposing an example to illustrate this use case, you can check out the code (Python) in this PR: https://github.com/stack-of-tasks/pinocchio/pull/1911/files

import numpy as np
from robot_descriptions.loaders.pinocchio import load_robot_description  # pip install robot_descriptions

robot = load_robot_description("upkie_description")
model = robot.model

# Joint placement offsets compared to the leg length
# We know from the robot description that both are along the local y-axis
known_offsets = {"knee": 0.072, "wheel": 0.065}

new_length = 0.3  # in meters, length for both femur and tibia links
for side in ("left", "right"):
    for joint in ("knee", "wheel"):
        joint_id = model.getJointId(f"{side}_{joint}")
        model.jointPlacements[joint_id].translation[1] = new_length - known_offsets[joint]