Archived

This topic is now archived and is closed to further replies.

gandalf3373

inertia tensor

Recommended Posts

gandalf3373    122
i am looking for a little help or maybe a more clear explanation about the inertia tensor. i am doing a simulation with some cubes and i fully understand how the moment of inertia about each axis is calculated and such. however i am confused with the tensor. what i think i understand is that i do not necessarily need the tensor represented as a 3 by 3 matrix to correctly represent the moment of inertia if my rigid body is completley symmetrical about each axis of rotation. like my cube, or a sphere or a rectangle. if the body is completely symmetrical about each axis in local or object space than i can represent its moment of inertia in my simulation as simply a vector, or in this case as a simple scalar actually since for the cube this value would be the same for each component (x,y,z). is this correct then??

Share this post


Link to post
Share on other sites
bpj1138    113
No, because you have to do a matrix multiply with the inertia tensor, when you''re doing the similarity transform (R * Ib * R''). R being the orientation, Ib inertia tensor in body coordinates, and R'' is the transpose of the orientation. You could store Ib as a vector and write a special multiply routine for it.

Share this post


Link to post
Share on other sites
oliii    2196
the vector would represent the diagonal of the matrix, but you''d be better off using a matrix straight away. The inertia tensor would be a diagonal matrix in object space, but not in world space.

Share this post


Link to post
Share on other sites