Warning: You are not logged in. Your IP address will be publicly visible if you make any edits. If you log in or create an account, your edits will be attributed to your username, along with other benefits.
The edit can be undone.
Please check the comparison below to verify that this is what you want to do, and then save the changes below to finish undoing the edit.
Latest revision |
Your text |
Line 284: |
Line 284: |
| * the code for the grid is here https://github.com/elf11/brlcad/blob/master/simulate.patch | | * the code for the grid is here https://github.com/elf11/brlcad/blob/master/simulate.patch |
| * made a comparison with what bullet returns https://docs.google.com/spreadsheet/ccc?key=0AsmiuNTm7UBZdG92bV9veW1GYk8xRS1zOWpMX0ltNmc#gid=0 | | * made a comparison with what bullet returns https://docs.google.com/spreadsheet/ccc?key=0AsmiuNTm7UBZdG92bV9veW1GYk8xRS1zOWpMX0ltNmc#gid=0 |
− |
| |
− | 21st October
| |
− |
| |
− | * tried to replicate the bullet standalone behaviour in brlcad, for that starting with getting the correct normals, started hardcoding them in the processCollision function from the simulationalgo.cpp file; last time the normals we got were (0,0,-1) we need (0, 0, 1) since the normals are from the ground plane to the cube and we want to replicate the standalone; it didn't work cause after hardcoding them we get only the first contact point, after that first contact point we don't get any more contacts between the 2 bodies which is obv wrong since the cube still falls through the gp
| |
− | * when the first contact is being processed then no more contacts are being added, but the cube is obv falling downwards since the depth grows
| |
− | * then tried to change the normals from the raytracing part, simrt.c file, the traverse_zray_lists() where we set up the sim_manifold strucure; it didn't work for the same reasons as before
| |
− | * the problem is that bullet keeps an array of manifolds and points for those manifolds, the m_manifoldPtr a pointer to this array (processCollision() function in the mged files and in the standalone bullet simulation) it gets cleared and updated every step;
| |
− | * last year student implemented its own structures for manifolds, never using the ones provided by bullet the sim_manifold struct in simulate.h; the difference is that in our processCollision function we work on 2 different structures, the bullet structure and our structure but bullet never gets the info from our structure past the first contact
| |
− |
| |
− | TO DO : change all the sim_manifold struct uses with the native bullet manifold structure or find a way for our particular structure to be compatible with bullet's structure
| |
| | | |
| TO DO : find a way to pick 4 points out of the 36 generated that cover the maximum area of the object | | TO DO : find a way to pick 4 points out of the 36 generated that cover the maximum area of the object |