massMatrix and inverseDynamics are something wrong....( Am I using wrong?)
16 views (last 30 days)
I've tested it by simple 2 DOF arm.
Then, I think massMatrix and inverseDynamics have some bugs..
In a 2 DOF arm, massMatrix gave me wrong inertia matrix...
(each length 1 [m], CoM : 0.5 [m], I = 0....... dynamics equations are already provided in the robotics textbook.)
InverseDynamics : summing inertia torque and coriolis torque and gravity torque may be something mistakes. Moreover, I can not believe calculated inertia torque values.
I am wondering if everyone is using mathworks's robot dynamics functions without problems.
Related documents need to be presented in more detail.
Boris Blagojevic on 22 Jun 2020
i encountered a similiar problem. The Mass Matrix of a 5-DoF Robot, calculated with Lagrange and Newton-Euler manually, doesn't match the result of the built-in massMatrix function. It appears to have a constant offset for all joint configurations for some entries, while others are equal to the analytically derived values.
Gravity Torque and Velocity Product seem to be fine at least.
This is a serious problem with the ROS-Toolbox and requires a response from the staff, if you ask me. I won't trust this Toolbox any longer, until this issue is fixed.