calculate inertia matrix by robotics toolbox

[Thắng Lê]

hello,

i have a problem when i use the dynamic function in robotics toolboxs.

i wrote a program in matlab to calculate inertia matrix. in order to check it, i use same parameter like mdl_puma560 in robotics toolbox but the results are difference. I dont know what wrong.I tried with the coriolis and gravity matrix. Only gravity matrix give result same. here is the formulation and code  i used to calculate inertia matrix 

[Erik van Oene]

Sorry, but I don't have the symbolic toolbox so I cannot help you in detail.

Can you specify the difference of the results? Is the whole matrix different or only some values?

Erik

Op maandag 18 september 2017 14:43:18 UTC+2 schreef Thắng Lê:

[Thắng Lê]

sorry about send wrong email.

1st, with this formulation the results quite difference from RTB, special is diagonal of inertial matrix and not logic (2 first values M11 and M22 are too small) then i left it. 

2nd, I tried a new formulation, the results is more positive. I compared it to RTB and ROS_KDL (orocos_kdl library). it same as ROS_KDL but still difference RTB. I dont know what formulation did RTB use or somethings were added in RTB because i only use inertia matrix, mass and center of mass, not sure if another parameters(gear ratio, frictions, gear inertia,..) effect these elements   

Vào 22:53:49 UTC+7 Thứ Sáu, ngày 22 tháng 9 năm 2017, Erik van Oene đã viết:

[Peter Corke]

The mdl_puma560 model is quite comprehensive, it includes friction and motor armature inertia.  Setting all the friction parameters and gear ratios to zero should then match the assumptions you have made.

Might be better to compare results for a simpler 2 or 3 link model also.

Peter