Rigid body system #10
Conversation
panchagil
commented
Jun 7, 2019
panchagil
commented
- Moved rigid bodies to physics/ folder
- added RigidBodySystem that keeps list of RB.
- Moved solvers to solvers/ folder
- Removed (a lot of)unused code
- moved RigidBody class to namespace ccd::opt - made RigidBody attributes private to ensure consistency of CM
- moved solvers to solver/ - moved rigid_body to physics/ - added RigidBodySystem that has a list of rigid bodies - handles the mapping of local to global coordinates
- Added get/set of RB velocities - Added method to clear RBSystem
| total_edge_length += edge_length; | ||
| } | ||
| Eigen::VectorXd cm; | ||
| cm = (vertex_masses.asDiagonal() * vertices).colwise().sum() |
There was a problem hiding this comment.
It would be nice to have a function to build the mass matrix. This function would then call the mass matrix function. The function would be
double build_mass_matrix(const Eigen::MatrixXd& vertices, const Eigen::MatrixXi& edges, Eigen::SparseMatrix<double>& mass_matrix)where the function return the total edge length.
There was a problem hiding this comment.
ok. I can make the build_mass_matrix. Is the total length the same as mass_matrix.sum()? If that is the case, I'll rather make the return void.
There was a problem hiding this comment.
Yeah, is the same as mass_matrix.sum(), so returning nothing makes more sense.
| = hvertices * Tm.transpose(); | ||
|
|
||
| Eigen::MatrixXd displacements; | ||
| displacements = tvertices.rowwise().hnormalized(); |
There was a problem hiding this comment.
I will convert this to not use homogeneous coordinates after we accept this pull request. It makes taking the gradient and hessian.
src/physics/rigid_body_system.hpp
Outdated
|
|
||
| Eigen::MatrixX2d world_vertices() const | ||
| { | ||
| return vertices.rowwise() + position.transpose(); |
There was a problem hiding this comment.
Why is this definition not in the cpp file?
There was a problem hiding this comment.
I've seen short functions defined on the .hpp file, idk why. I can move it to the cpp.
| void assemble_displacements(); | ||
| void assemble_displacements( | ||
| const Eigen::VectorXd& v, Eigen::MatrixXd& u); | ||
| void add_rigid_body(RigidBody rb) { rigid_bodies.push_back(rb); } |
There was a problem hiding this comment.
Should this function call assemble or do something to update the fields?
There was a problem hiding this comment.
I thought it would be expensive if we need to add many rigid bodies.
There was a problem hiding this comment.
Oh okay, so when is assemble called?
There was a problem hiding this comment.
after you add them all. Look at state.cpp update_fields_from_rigid_bodies.
| @@ -1,6 +1,6 @@ | |||
| // Functions for optimizing functions. | |||
| // Includes Newton's method with and without constraints. | |||
| #include <opt/newtons_method.hpp> | |||
| #include "newtons_method.hpp" | |||
There was a problem hiding this comment.
Why use quotes here when we mostly use <solvers/newton_method.hpp>?
There was a problem hiding this comment.
I've been using quotes when is its own .hpp, mostly because that is the template of my IDE.
