torque/force sensor problem

[Swk6666]

Intro

Hi!

I am a graduate student at beijing institute of technology, I use MuJoCo for my research on robot simulation.

My setup

mujoco version is 3.3.0 ,python3.10 ,mac os15

My question

I have created a scenario with a dual 7-degree-of-freedom robotic arm grasping an object to form a closed-chain constraint, as shown in the figure. The connection between the robotic arm and the object is established through two weld constraints. I placed force sensors at joint1_7 and joint2_7, expecting to obtain 6-dimensional force data. When I set self.data.eq_active = np.array([0,0]), the z-direction torque value from the torque sensor equals the value from data.qfrc_actuator, and my RNE program aligns well with the torque output from MuJoCo (self.data.qfrc_actuator[0:14]). However, when I set self.data.eq_active = np.array([1,1]), np.array([1,0]), or np.array([0,1]), the z-direction torque value from the torque sensor no longer matches the value from data.qfrc_actuator, and the results no longer align. I want to know what might be going wrong. Is it possible that the sensors are not fully measuring the forces? Below are the relevant settings from my XML file. The corresponding files are included in the zip file, and you can reproduce the issue by running main.py.


Minimal model and/or code that explain my question

github_issue.zip

Confirmations

• I searched the latest documentation thoroughly before posting.
• I searched previous Issues and Discussions, I am certain this has not been raised before.

[Balint-H]

Hello!

From the documentation of the force sensor:

This element creates a 3-axis force sensor. The sensor outputs three numbers, which are the interaction force between a child and a parent body, expressed in the site frame defining the sensor. The convention is that the site is attached to the child body, and the force points from the child towards the parent. The computation here takes into account all forces acting on the system, including contacts as well as external perturbations.

With the torque sensor being similar applied in the parent's frame

Do I understand correctly you are interested in the force between the object and the end-effector, and not the end-effector and its parent (expressed in the corresponding frame)? If yes, then that could lead to the differences in expected and observed readings.

[Swk6666]

That's not the case. On the contrary, I am interested in the 6-dimensional force between the end effector(link1-7) and its parent（link1-6）, not the constraint force between the end effector(link1-7) and an object (object) generated through a weld constraint.