inertial matrix is not symmetric

[huiwenzhang]

Hi, the inertial matrix should be symmetric, but a lot of demos in this repo have non-symmetric inertial matrix update, pls check.

[wissem01chiha]

Hi @huiwenzhang , thanks very very much for your feedback, and sorry for late replay, for clarification, do you refer for the inertial matrix writing in the URDF of the robot or those computed in the .npy autogenerated files folder ? , for the urdf parameters , they are extracted the same of the official kinova URDF file , for the computed they may not be symmetric , and this is a problem with parameters physically Consistency which i hadn't the time to focus more, have you any fixing proposals ? i am welcoming modification

[huiwenzhang]

Hi, it is the computed inertial matrix. I am trying to run the parameters identification demo. To my understanding, the program will use the linear minimum parameter set the actuation trajectory to estimate the parameters in Y matrix. As for the inertial matrix, I think it is always should be symmetric, which implies only 6 of 9 elements in inertial matrix should be optimized. I checked the log, the matrix is always non-symmetric. I have't dive deep into the code.

[wissem01chiha]

@huiwenzhang Hi, yes, I understand. Which script are you trying to run first? I will check the solver being used and the objective function it is trying to minimize.