I would like to express some doubts regarding the accuracy of some masses and inertias in the xacro files. As I am simulating a robotnik kairos in ROS using Rviz and Gazebo, I am relying on an atleast semi-stable gazebo simulation. During tests, I have noticed that the mobile manipulator tends to tip over much more easily than I expected, which lead me to investigating the masses and inertias of the links in the urdf. In the entire urdf tree, two links have caught my attention: ${prefix}base_link and ${prefix}_wheel_link, of which I have listed the important parts of the URDF below (although I have provided urls from the melodic branch, it seems to me that the same code snippets are used in all branches):
As a minor point, there seems to be a discrepancy in total mass, because 125+4*6.5=151, which is 11kg more than in the datasheet which amounts roughly to a relative error of 10%. However I am much more concerned about the axial moments of inertia of the base_link, specifically why ixx (1.4) is four times lower than iyy (6.9) or izz (6.1). I have done some rough calculations with assumptions about the distribution of mass (ixx=6.4, iyy=7.7, izz=5,3), which resulted in a significantly larger ixx while my iyy and izz seem to be around the official values in the URDF.
Both my visual intuition (summit base dimensions are roughly in the same order of magnitude) and the rough mathematical estimate are telling me that the values are not correct. Furthermore, I am suspicious that all summit bases have the same inertia values in the URDF.
Hi
I would like to express some doubts regarding the accuracy of some masses and inertias in the xacro files. As I am simulating a robotnik kairos in ROS using Rviz and Gazebo, I am relying on an atleast semi-stable gazebo simulation. During tests, I have noticed that the mobile manipulator tends to tip over much more easily than I expected, which lead me to investigating the masses and inertias of the links in the urdf. In the entire urdf tree, two links have caught my attention: ${prefix}base_link and ${prefix}_wheel_link, of which I have listed the important parts of the URDF below (although I have provided urls from the melodic branch, it seems to me that the same code snippets are used in all branches):
source: https://github.com/RobotnikAutomation/summit_xl_common/blob/melodic-master/summit_xl_description/urdf/bases/summit_xls_base.urdf.xacro
source: https://github.com/RobotnikAutomation/summit_xl_common/blob/melodic-master/summit_xl_description/urdf/wheels/omni_wheel.urdf.xacro
If I am not mistaken, the data sheet for the robot I am working on (SXLSK-190801) and which I have used my calculations is: https://robotnik.eu/wp-content/uploads/2020/06/DATASHEET_RB-KAIROS-16_EN.pdf
As a minor point, there seems to be a discrepancy in total mass, because 125+4*6.5=151, which is 11kg more than in the datasheet which amounts roughly to a relative error of 10%. However I am much more concerned about the axial moments of inertia of the base_link, specifically why ixx (1.4) is four times lower than iyy (6.9) or izz (6.1). I have done some rough calculations with assumptions about the distribution of mass (ixx=6.4, iyy=7.7, izz=5,3), which resulted in a significantly larger ixx while my iyy and izz seem to be around the official values in the URDF.
Both my visual intuition (summit base dimensions are roughly in the same order of magnitude) and the rough mathematical estimate are telling me that the values are not correct. Furthermore, I am suspicious that all summit bases have the same inertia values in the URDF.